Glossary

Acronyms, jargon and slang from chat, discussions and interview write-ups. Every term has its own link, so you can point someone at /glossary#tle.

#

1p3a1point3acres (一亩三分地): A Chinese-language forum about studying and working in the US, and the largest source of interview experience posts. See 面经.
面经miànjīng, interview experience: A post describing the questions from a recent interview. The word is Chinese; most of these posts appear on 1point3acres. See 1p3a.

A

ACAccepted: The submission produced the expected output on every test case within the time and memory limits.
Amortized: The average cost per operation over a long sequence of operations. Appending to a dynamic array is amortized O(1): an occasional resize is O(n), but it happens rarely enough that the average stays constant.
AtCoder: A Japanese contest site. ABC (Beginner), ARC (Regular) and AGC (Grand) are its contest tiers, easiest to hardest.

B

Backtracking: Building candidates one choice at a time and undoing the last choice when it cannot lead to a valid answer. Used for permutations, subsets and constraint puzzles.
Bar raiser: An interviewer from outside the hiring team whose job is to hold the company-wide hiring standard. Originally an Amazon term.
Base / bonus / sign-on: The cash parts of an offer: yearly salary, a yearly performance bonus, and a one-time sign-on payment that usually has to be repaid if you leave within a year. See TC (Total Compensation).
Beats 99%: The runtime percentile LeetCode shows after an accepted submission, compared with other accepted submissions in the same language. It varies with server load, so the same code can land far apart on two runs.
Behavioral: An interview about past work and how the candidate handled specific situations. See STAR.
BFSBreadth-First Search: Visits a graph or tree level by level using a queue. Finds shortest paths when every edge has the same weight. See DFS.
Big O: Notation for how running time or memory grows with input size, ignoring constant factors. O(n) grows linearly, O(n²) quadratically, O(log n) logarithmically.
BITBinary Indexed Tree: A compact array structure for prefix sums with point updates, both in O(log n). Also called a Fenwick tree. See Segment tree.
Bitmask: An integer whose bits mark which items of a small set are chosen. "Bitmask DP" uses these integers as DP states.
Blind: An anonymous forum for tech employees, verified by work email. Posts show the author's company, which is why comp threads there carry weight.
Blind 75: A widely shared list of 75 LeetCode problems covering the core interview patterns, first posted on the Blind forum. See NeetCode 150.
Boomerang: Someone who leaves a company and later returns to it.
Brute force: A solution that tries every possibility. Usually correct and too slow for the full constraints.
BSTBinary Search Tree: A binary tree where every node's left subtree holds smaller values and its right subtree larger ones.

C

CECompilation Error: The code did not compile, so no test case ran. For interpreted languages this covers syntax errors caught before execution.
CFCodeforces: The most active competitive programming site. Contests are split by rating into divisions: Div. 1 for the strongest, Div. 2 to Div. 4 for everyone else.
Cheese: A solution that passes by exploiting weak tests or loose limits instead of solving the intended problem.
Cliff: The period before any stock vests, commonly one year. Leaving before the cliff forfeits the whole grant. See RSU.
Competing offer: An offer from another company, used as leverage to negotiate a higher one.
Cooked: In a hopeless position: "got a hard in the phone screen, I'm cooked".
Cooldown: The waiting period, typically six months to a year, before a company lets a rejected candidate apply again.
CPCompetitive Programming: Solving algorithm problems against the clock in contests, scored by the judge.

D

DAGDirected Acyclic Graph: A directed graph with no cycles.
Debrief: The meeting after an interview loop where the interviewers compare feedback and reach a recommendation. See Packet.
DFSDepth-First Search: Follows one path as deep as it goes before backtracking, using recursion or an explicit stack. See BFS, Backtracking.
Dijkstra: The shortest-path algorithm for graphs with non-negative edge weights, usually run with a priority queue. See PQ.
Discuss: LeetCode's forum, and by extension the community solutions posted under a problem. Interview experience posts live there too.
Down-level / up-level: Being offered a lower (or higher) level than the one you interviewed for or hold now. See L3, L4, E5 ….
DPDynamic Programming: Solving a problem by combining answers to overlapping subproblems, each computed once and stored. See Memo, Tabulation.
Dry run: Tracing code by hand on a small example, step by step, to check it before running it.
DSAData Structures and Algorithms: The subject area interview problems are drawn from.
DSUDisjoint Set Union: A structure that tracks which elements belong to the same group and merges groups in near-constant time. Also called Union-Find.

E

E / M / HEasy / Medium / Hard: Problem difficulty tiers. "Two mediums in 45 minutes" is a common interview format.
Edge case: An input at the boundary of the constraints: an empty array, a single element, all duplicates, the maximum value.
Editorial: The official written solution to a problem, usually covering several approaches from brute force to optimal.
EMEngineering Manager: Manages a team of engineers. A TL (Tech Lead) leads the technical work without managing people; a TLM (Tech Lead Manager) does both. See IC.
Exploding offer: An offer with a short deadline to accept.

F

FAANGFacebook, Amazon, Apple, Netflix, Google: Shorthand for the largest, best-paying tech companies. MAANG swaps in Meta; "big tech" means roughly the same.
Follow-up: A harder variant the interviewer asks after the original problem is solved, such as a lower complexity target or a streaming input.
Frequency: A Premium column that ranks a company's tagged problems by how often they have been reported recently. See Tagged.
Front-loaded / back-loaded: How a stock grant is spread over its vesting years. Amazon's 5/15/40/40 is the best-known back-loaded schedule; front-loaded grants pay out more in the first years. See RSU.

G

GGGood Game: Said at the end of a contest or session. "GG" on its own can also mean something is over and lost.
Ghosted: Stopped hearing back from a recruiter or company without a decision.
Goated: Excellent. From GOAT, "greatest of all time".
Golden handcuffs: Unvested stock or pay that makes leaving too expensive, even for someone who wants to go.
Googleyness: Google's name for the collaboration and values part of its assessment. Other companies call the same thing culture fit.
Greedy: Making the locally best choice at each step without revisiting it. Correct only for problems where that choice can be proven safe.
Grind: Steady, high-volume practice, as in "grinding LeetCode". See Grind 75.
Grind 75: An updated study list from the author of the Blind 75. It adjusts to how many weeks are left and extends to 169 problems. See Blind 75.

H

H-1B: The main US work visa for specialty occupations. The employer sponsors it, new visas are allocated by an annual lottery, and it lasts three years, renewable once. See OPT / STEM OPT, PERM.
Hack: On Codeforces, submitting an input that makes another contestant's accepted solution fail. A hacked solution loses its points.
HCHiring Committee, or Headcount: Two meanings. A hiring committee reviews interview feedback and makes the hire decision (Google is the best-known example). Headcount is an approved open position, as in "the team lost HC".
Hidden tests: Test cases that run on Submit but are not shown beforehand. Run checks only the sample cases; a solution that passes the samples can still fail a hidden one.
Hints: Help the interviewer gave during the round. Feedback records how many were needed and how big they were, and the score takes them into account.
Hiring freeze: A pause on filling open positions. Candidates already mid-loop can be dropped without a decision.

I

ICIndividual Contributor: An engineer who does not manage people. Most companies run parallel IC and management ladders. See EM, Staff / Principal.
ICPCInternational Collegiate Programming Contest: A team contest for university students: three people share one computer.
IIRC / AFAIKIf I Recall Correctly / As Far As I Know: Marks a statement the writer is fairly but not completely sure of.
In-place: Modifying the input directly instead of building a copy, using O(1) extra space apart from a few variables.
IOIInternational Olympiad in Informatics: The international programming olympiad for secondary-school students.
It's so over / we're so back: Mock despair and mock triumph, often about the same thing a day apart.

J

Job hopping: Changing companies every year or two, usually for a raise. Switching tends to raise pay faster than staying.

K

KMPKnuth–Morris–Pratt: A string-matching algorithm that finds every occurrence of a pattern in a text in linear time.

L

L3, L4, E5 …: Company-specific seniority levels. L3 at Google and E3 at Meta are entry level; each number up is one step of seniority.
LCLeetCode: The largest interview-practice site. Problems are often referred to by number, as in "LC 200".
LC Premium: LeetCode's paid subscription. It unlocks company tags, frequency sorting, locked problems and most editorials. See Tagged, Frequency.
LCALowest Common Ancestor: The deepest node in a tree that has both of two given nodes as descendants.
LCSLongest Common Subsequence: A classic DP problem: the longest sequence that appears, in order, in both of two strings.
levels.fyi: A site of crowd-sourced compensation data by company, level and location. The usual reference in offer negotiations.
LGTMLooks Good To Me: Approval in a code review.
LISLongest Increasing Subsequence: A classic DP problem, and the name of its technique: the longest subsequence whose elements strictly increase.
LLDLow-Level Design: An interview about designing classes and interfaces for a small system, such as a parking lot. HLD (High-Level Design) is another name for system design. See SD.
Love Babbar 450: A list of 450 DSA problems compiled by YouTuber Love Babbar, popular in India.
Lowball: An offer well below the going rate for the level and location.
LPLeadership Principles: Amazon's list of values. Its behavioral questions are organised around them.

M

MemoMemoization: Caching a recursive function's results by argument so each subproblem is computed once. Also called top-down DP. See DP.
Mid: Mediocre. Not to be confused with a medium-difficulty problem.
MLEMemory Limit Exceeded: The program used more memory than the limit allows.
MockMock interview: A practice interview run like the real thing, with someone else asking the questions.
Monotonic stack: A stack kept in increasing or decreasing order by popping elements that would break the order. Finds the next greater or smaller element for every position in O(n).
MSTMinimum Spanning Tree: The cheapest set of edges that connects every node of a weighted graph. Found with Kruskal's or Prim's algorithm.

N

NeetCode 150: An expanded study list of 150 problems built on the Blind 75, organised by pattern. See Blind 75.
NGNew Grad: An entry-level role for recent graduates.

O

OAOnline Assessment: A timed coding test, usually automated and taken alone, that screens candidates before any live interview.
Off-by-one: A bug where a loop or index is one step too far or too short, such as < where <= was needed.
OLEOutput Limit Exceeded: The program printed more output than the judge accepts. Leftover debug prints inside a loop are the usual cause.
One-shot: Solved on the first attempt with no failed submissions.
Onsite: The final round: several back-to-back interviews covering coding, system design and behavioral questions. The name stuck even though most are now held over video. See VO.
OPT / STEM OPTOptional Practical Training: Work authorization for F-1 students in the US after graduation: 12 months, plus a 24-month extension for STEM degrees. See H-1B.
Optimal: In interview talk, the solution with the best known complexity for the problem, as opposed to the brute force.

P

Packet: Everything a hiring committee reads about a candidate: interview feedback, résumé and recruiter notes. See HC.
PERM: The labor certification step of an employer-sponsored US green card. "Does the company sponsor PERM?" is a common offer question. See H-1B.
Phone screen: The first live technical interview, usually one problem in a shared editor over a video call.
PIPPerformance Improvement Plan: A formal warning with targets to meet in a fixed period, often a precursor to being let go.
POTDProblem of the Day: The daily problem a site features for everyone. Solving it every day keeps a streak going.
PQPriority Queue: A collection that always hands back the smallest (or largest) element first. Usually implemented as a binary heap.
Prefix sum: An array where entry i holds the sum of the first i elements, so any range sum is one subtraction.

R

Rating: A number that rises and falls with contest results, like chess Elo. Sites attach titles to rating bands, such as Codeforces' Specialist, Expert and Master or LeetCode's Knight and Guardian.
RERuntime Error: The program crashed or exited with a non-zero status: an index out of range, a null dereference, a stack overflow from deep recursion, an uncaught exception.
Refresher: An additional stock grant given after hire, usually yearly and tied to performance, so that pay does not drop once the first grant has vested. See RSU.
Reneged / rescinded: An accepted offer withdrawn: reneged by the candidate, rescinded by the company.
Return offer: A full-time offer made to an intern at the end of the internship.
RIFReduction In Force: The formal term for layoffs.
RSURestricted Stock Unit: Company stock granted as pay, delivered in portions over a vesting schedule, typically four years.
RTOReturn To Office: A policy requiring remote or hybrid staff to come back to the office for a set number of days.
Rubber ducking: Explaining code line by line, to a person or an object, until the bug becomes obvious.
Rubric: The written criteria a company scores each round against, usually covering problem solving, coding, communication and testing.

S

SCSpace Complexity: How extra memory grows with input size. Usually excludes the input itself and, by convention, the output. See TC (Time Complexity).
SDSystem Design: An interview where the candidate designs a large system, such as a URL shortener or a chat service, and discusses the trade-offs. See LLD.
Sean Prashad list: A free list of LeetCode problems grouped by pattern, built on the Blind 75. See Blind 75.
Segment tree: A tree over an array that answers range queries (sum, minimum, maximum) and applies updates in O(log n). See BIT.
Shadow / reverse shadow: The stages of interviewer training. A new interviewer first watches an experienced one run the round (shadow), then runs it while the experienced one watches (reverse shadow).
Signal: Evidence an interviewer collected about a skill: "strong signal on coding, not enough signal on design". A round that ends without it is scored as if the skill were missing.
Skill issue: Teasing: the problem was the person, not the tool or the question.
Sliding window: A contiguous range whose edges move forward through an array or string while a running summary of its contents is kept up to date.
Stack ranking: Performance reviews where a fixed share of each team must be rated at the bottom, whatever their actual performance.
Staff / Principal: The IC levels above Senior. Staff engineers lead work across teams; Principal engineers work across an organisation. See IC, L3, L4, E5 ….
STARSituation, Task, Action, Result: A common structure for answering behavioral questions.
Striver's SDE Sheet: A widely used problem list from the takeUforward YouTube channel, popular in India.
Strong Hire … Strong No Hire: The rating scale interviewers use to score a round. The usual steps are Strong No Hire, No Hire, Lean No Hire, Lean Hire, Hire and Strong Hire. See Debrief.
SWESoftware Engineer: The generic job title. Amazon and some others use SDE (Software Development Engineer).

T

Tabulation: Filling a DP table iteratively from the smallest subproblems up. Also called bottom-up DP. See DP.
Tagged: Problems a site lists as reported from a specific company's interviews, as in "Google tagged". Grinding tagged problems is a common last step before an interview.
TCTime Complexity: How running time grows with input size, written in Big O. Often paired with SC in solution posts: "TC O(n), SC O(1)". See SC, TC (Total Compensation), TC (The Candidate).
TCTotal Compensation: Base salary plus bonus plus stock, per year. In offer posts "TC" means this, not time complexity. See TC (Time Complexity), TC (The Candidate), RSU.
TCThe Candidate: Shorthand for the person being interviewed, used in interview feedback and mock-interview notes: "TC asked good clarifying questions." See TC (Time Complexity), TC (Total Compensation).
TC or GTFO: Blind's standing demand that anyone offering an opinion first post their total compensation. See TC (Total Compensation).
Team match: The stage after passing interviews where a candidate talks to teams with open positions before an offer is final.
TILToday I Learned: Introduces something the writer just found out.
TLETime Limit Exceeded: A test case ran longer than the time limit. Usually the algorithm is too slow for the largest inputs, but an infinite loop produces the same verdict. See Big O.
Topo sortTopological sort: An ordering of a directed acyclic graph's nodes where every edge points forward. Used for dependency and scheduling problems. See DAG.
Trie: A tree of characters where each path from the root spells a prefix. Used for prefix lookups and word searches. Pronounced "try".
Two pointers: Two indices moving through an array or string, often from both ends or at different speeds.

U

Upsolve: Solving, after a contest ends, the problems you could not finish during it.
URAUnregretted Attrition: Amazon's term for employees leaving whom the company is content to lose. On Blind it usually refers to reported targets for how many that should be. See Stack ranking, PIP.
USACOUSA Computing Olympiad: A contest for pre-college students with Bronze, Silver, Gold and Platinum divisions.

V

Verbal / written offer: A verbal offer is the recruiter telling you the numbers; the written offer is the signed document. Only the written one is binding.
Virtual contest: A past contest replayed under its original time limit, as practice.
VOVirtual Onsite: An onsite held over video. See Onsite.

W

W / LWin / Loss: A good or bad outcome: "passed the OA, W".
WAWrong Answer: The program ran to completion but its output differed from the expected output on at least one test case.
Weekly / Biweekly: LeetCode's contests: the weekly runs every Sunday, the biweekly every other Saturday. Each has four problems and 90 minutes.
WLBWork-Life Balance: How much of your life the job leaves alone. Team-by-team WLB is one of the most-asked questions on Blind.

Y

YOEYears of Experience: Professional experience, as in "3 YOE".

Glossary

Acronyms, jargon and slang from chat, discussions and interview write-ups. Every term has its own link, so you can point someone at /glossary#tle.

#

1p3a1point3acres (一亩三分地): A Chinese-language forum about studying and working in the US, and the largest source of interview experience posts. See 面经.
面经miànjīng, interview experience: A post describing the questions from a recent interview. The word is Chinese; most of these posts appear on 1point3acres. See 1p3a.

A

ACAccepted: The submission produced the expected output on every test case within the time and memory limits.
Amortized: The average cost per operation over a long sequence of operations. Appending to a dynamic array is amortized O(1): an occasional resize is O(n), but it happens rarely enough that the average stays constant.
AtCoder: A Japanese contest site. ABC (Beginner), ARC (Regular) and AGC (Grand) are its contest tiers, easiest to hardest.

B

Backtracking: Building candidates one choice at a time and undoing the last choice when it cannot lead to a valid answer. Used for permutations, subsets and constraint puzzles.
Bar raiser: An interviewer from outside the hiring team whose job is to hold the company-wide hiring standard. Originally an Amazon term.
Base / bonus / sign-on: The cash parts of an offer: yearly salary, a yearly performance bonus, and a one-time sign-on payment that usually has to be repaid if you leave within a year. See TC (Total Compensation).
Beats 99%: The runtime percentile LeetCode shows after an accepted submission, compared with other accepted submissions in the same language. It varies with server load, so the same code can land far apart on two runs.
Behavioral: An interview about past work and how the candidate handled specific situations. See STAR.
BFSBreadth-First Search: Visits a graph or tree level by level using a queue. Finds shortest paths when every edge has the same weight. See DFS.
Big O: Notation for how running time or memory grows with input size, ignoring constant factors. O(n) grows linearly, O(n²) quadratically, O(log n) logarithmically.
BITBinary Indexed Tree: A compact array structure for prefix sums with point updates, both in O(log n). Also called a Fenwick tree. See Segment tree.
Bitmask: An integer whose bits mark which items of a small set are chosen. "Bitmask DP" uses these integers as DP states.
Blind: An anonymous forum for tech employees, verified by work email. Posts show the author's company, which is why comp threads there carry weight.
Blind 75: A widely shared list of 75 LeetCode problems covering the core interview patterns, first posted on the Blind forum. See NeetCode 150.
Boomerang: Someone who leaves a company and later returns to it.
Brute force: A solution that tries every possibility. Usually correct and too slow for the full constraints.
BSTBinary Search Tree: A binary tree where every node's left subtree holds smaller values and its right subtree larger ones.

C

CECompilation Error: The code did not compile, so no test case ran. For interpreted languages this covers syntax errors caught before execution.
CFCodeforces: The most active competitive programming site. Contests are split by rating into divisions: Div. 1 for the strongest, Div. 2 to Div. 4 for everyone else.
Cheese: A solution that passes by exploiting weak tests or loose limits instead of solving the intended problem.
Cliff: The period before any stock vests, commonly one year. Leaving before the cliff forfeits the whole grant. See RSU.
Competing offer: An offer from another company, used as leverage to negotiate a higher one.
Cooked: In a hopeless position: "got a hard in the phone screen, I'm cooked".
Cooldown: The waiting period, typically six months to a year, before a company lets a rejected candidate apply again.
CPCompetitive Programming: Solving algorithm problems against the clock in contests, scored by the judge.

D

DAGDirected Acyclic Graph: A directed graph with no cycles.
Debrief: The meeting after an interview loop where the interviewers compare feedback and reach a recommendation. See Packet.
DFSDepth-First Search: Follows one path as deep as it goes before backtracking, using recursion or an explicit stack. See BFS, Backtracking.
Dijkstra: The shortest-path algorithm for graphs with non-negative edge weights, usually run with a priority queue. See PQ.
Discuss: LeetCode's forum, and by extension the community solutions posted under a problem. Interview experience posts live there too.
Down-level / up-level: Being offered a lower (or higher) level than the one you interviewed for or hold now. See L3, L4, E5 ….
DPDynamic Programming: Solving a problem by combining answers to overlapping subproblems, each computed once and stored. See Memo, Tabulation.
Dry run: Tracing code by hand on a small example, step by step, to check it before running it.
DSAData Structures and Algorithms: The subject area interview problems are drawn from.
DSUDisjoint Set Union: A structure that tracks which elements belong to the same group and merges groups in near-constant time. Also called Union-Find.

E

E / M / HEasy / Medium / Hard: Problem difficulty tiers. "Two mediums in 45 minutes" is a common interview format.
Edge case: An input at the boundary of the constraints: an empty array, a single element, all duplicates, the maximum value.
Editorial: The official written solution to a problem, usually covering several approaches from brute force to optimal.
EMEngineering Manager: Manages a team of engineers. A TL (Tech Lead) leads the technical work without managing people; a TLM (Tech Lead Manager) does both. See IC.
Exploding offer: An offer with a short deadline to accept.

F

FAANGFacebook, Amazon, Apple, Netflix, Google: Shorthand for the largest, best-paying tech companies. MAANG swaps in Meta; "big tech" means roughly the same.
Follow-up: A harder variant the interviewer asks after the original problem is solved, such as a lower complexity target or a streaming input.
Frequency: A Premium column that ranks a company's tagged problems by how often they have been reported recently. See Tagged.
Front-loaded / back-loaded: How a stock grant is spread over its vesting years. Amazon's 5/15/40/40 is the best-known back-loaded schedule; front-loaded grants pay out more in the first years. See RSU.

G

GGGood Game: Said at the end of a contest or session. "GG" on its own can also mean something is over and lost.
Ghosted: Stopped hearing back from a recruiter or company without a decision.
Goated: Excellent. From GOAT, "greatest of all time".
Golden handcuffs: Unvested stock or pay that makes leaving too expensive, even for someone who wants to go.
Googleyness: Google's name for the collaboration and values part of its assessment. Other companies call the same thing culture fit.
Greedy: Making the locally best choice at each step without revisiting it. Correct only for problems where that choice can be proven safe.
Grind: Steady, high-volume practice, as in "grinding LeetCode". See Grind 75.
Grind 75: An updated study list from the author of the Blind 75. It adjusts to how many weeks are left and extends to 169 problems. See Blind 75.

H

H-1B: The main US work visa for specialty occupations. The employer sponsors it, new visas are allocated by an annual lottery, and it lasts three years, renewable once. See OPT / STEM OPT, PERM.
Hack: On Codeforces, submitting an input that makes another contestant's accepted solution fail. A hacked solution loses its points.
HCHiring Committee, or Headcount: Two meanings. A hiring committee reviews interview feedback and makes the hire decision (Google is the best-known example). Headcount is an approved open position, as in "the team lost HC".
Hidden tests: Test cases that run on Submit but are not shown beforehand. Run checks only the sample cases; a solution that passes the samples can still fail a hidden one.
Hints: Help the interviewer gave during the round. Feedback records how many were needed and how big they were, and the score takes them into account.
Hiring freeze: A pause on filling open positions. Candidates already mid-loop can be dropped without a decision.

I

ICIndividual Contributor: An engineer who does not manage people. Most companies run parallel IC and management ladders. See EM, Staff / Principal.
ICPCInternational Collegiate Programming Contest: A team contest for university students: three people share one computer.
IIRC / AFAIKIf I Recall Correctly / As Far As I Know: Marks a statement the writer is fairly but not completely sure of.
In-place: Modifying the input directly instead of building a copy, using O(1) extra space apart from a few variables.
IOIInternational Olympiad in Informatics: The international programming olympiad for secondary-school students.
It's so over / we're so back: Mock despair and mock triumph, often about the same thing a day apart.

J

Job hopping: Changing companies every year or two, usually for a raise. Switching tends to raise pay faster than staying.

K

KMPKnuth–Morris–Pratt: A string-matching algorithm that finds every occurrence of a pattern in a text in linear time.

L

L3, L4, E5 …: Company-specific seniority levels. L3 at Google and E3 at Meta are entry level; each number up is one step of seniority.
LCLeetCode: The largest interview-practice site. Problems are often referred to by number, as in "LC 200".
LC Premium: LeetCode's paid subscription. It unlocks company tags, frequency sorting, locked problems and most editorials. See Tagged, Frequency.
LCALowest Common Ancestor: The deepest node in a tree that has both of two given nodes as descendants.
LCSLongest Common Subsequence: A classic DP problem: the longest sequence that appears, in order, in both of two strings.
levels.fyi: A site of crowd-sourced compensation data by company, level and location. The usual reference in offer negotiations.
LGTMLooks Good To Me: Approval in a code review.
LISLongest Increasing Subsequence: A classic DP problem, and the name of its technique: the longest subsequence whose elements strictly increase.
LLDLow-Level Design: An interview about designing classes and interfaces for a small system, such as a parking lot. HLD (High-Level Design) is another name for system design. See SD.
Love Babbar 450: A list of 450 DSA problems compiled by YouTuber Love Babbar, popular in India.
Lowball: An offer well below the going rate for the level and location.
LPLeadership Principles: Amazon's list of values. Its behavioral questions are organised around them.

M

MemoMemoization: Caching a recursive function's results by argument so each subproblem is computed once. Also called top-down DP. See DP.
Mid: Mediocre. Not to be confused with a medium-difficulty problem.
MLEMemory Limit Exceeded: The program used more memory than the limit allows.
MockMock interview: A practice interview run like the real thing, with someone else asking the questions.
Monotonic stack: A stack kept in increasing or decreasing order by popping elements that would break the order. Finds the next greater or smaller element for every position in O(n).
MSTMinimum Spanning Tree: The cheapest set of edges that connects every node of a weighted graph. Found with Kruskal's or Prim's algorithm.

N

NeetCode 150: An expanded study list of 150 problems built on the Blind 75, organised by pattern. See Blind 75.
NGNew Grad: An entry-level role for recent graduates.

O

OAOnline Assessment: A timed coding test, usually automated and taken alone, that screens candidates before any live interview.
Off-by-one: A bug where a loop or index is one step too far or too short, such as < where <= was needed.
OLEOutput Limit Exceeded: The program printed more output than the judge accepts. Leftover debug prints inside a loop are the usual cause.
One-shot: Solved on the first attempt with no failed submissions.
Onsite: The final round: several back-to-back interviews covering coding, system design and behavioral questions. The name stuck even though most are now held over video. See VO.
OPT / STEM OPTOptional Practical Training: Work authorization for F-1 students in the US after graduation: 12 months, plus a 24-month extension for STEM degrees. See H-1B.
Optimal: In interview talk, the solution with the best known complexity for the problem, as opposed to the brute force.

P

Packet: Everything a hiring committee reads about a candidate: interview feedback, résumé and recruiter notes. See HC.
PERM: The labor certification step of an employer-sponsored US green card. "Does the company sponsor PERM?" is a common offer question. See H-1B.
Phone screen: The first live technical interview, usually one problem in a shared editor over a video call.
PIPPerformance Improvement Plan: A formal warning with targets to meet in a fixed period, often a precursor to being let go.
POTDProblem of the Day: The daily problem a site features for everyone. Solving it every day keeps a streak going.
PQPriority Queue: A collection that always hands back the smallest (or largest) element first. Usually implemented as a binary heap.
Prefix sum: An array where entry i holds the sum of the first i elements, so any range sum is one subtraction.

R

Rating: A number that rises and falls with contest results, like chess Elo. Sites attach titles to rating bands, such as Codeforces' Specialist, Expert and Master or LeetCode's Knight and Guardian.
RERuntime Error: The program crashed or exited with a non-zero status: an index out of range, a null dereference, a stack overflow from deep recursion, an uncaught exception.
Refresher: An additional stock grant given after hire, usually yearly and tied to performance, so that pay does not drop once the first grant has vested. See RSU.
Reneged / rescinded: An accepted offer withdrawn: reneged by the candidate, rescinded by the company.
Return offer: A full-time offer made to an intern at the end of the internship.
RIFReduction In Force: The formal term for layoffs.
RSURestricted Stock Unit: Company stock granted as pay, delivered in portions over a vesting schedule, typically four years.
RTOReturn To Office: A policy requiring remote or hybrid staff to come back to the office for a set number of days.
Rubber ducking: Explaining code line by line, to a person or an object, until the bug becomes obvious.
Rubric: The written criteria a company scores each round against, usually covering problem solving, coding, communication and testing.

S

SCSpace Complexity: How extra memory grows with input size. Usually excludes the input itself and, by convention, the output. See TC (Time Complexity).
SDSystem Design: An interview where the candidate designs a large system, such as a URL shortener or a chat service, and discusses the trade-offs. See LLD.
Sean Prashad list: A free list of LeetCode problems grouped by pattern, built on the Blind 75. See Blind 75.
Segment tree: A tree over an array that answers range queries (sum, minimum, maximum) and applies updates in O(log n). See BIT.
Shadow / reverse shadow: The stages of interviewer training. A new interviewer first watches an experienced one run the round (shadow), then runs it while the experienced one watches (reverse shadow).
Signal: Evidence an interviewer collected about a skill: "strong signal on coding, not enough signal on design". A round that ends without it is scored as if the skill were missing.
Skill issue: Teasing: the problem was the person, not the tool or the question.
Sliding window: A contiguous range whose edges move forward through an array or string while a running summary of its contents is kept up to date.
Stack ranking: Performance reviews where a fixed share of each team must be rated at the bottom, whatever their actual performance.
Staff / Principal: The IC levels above Senior. Staff engineers lead work across teams; Principal engineers work across an organisation. See IC, L3, L4, E5 ….
STARSituation, Task, Action, Result: A common structure for answering behavioral questions.
Striver's SDE Sheet: A widely used problem list from the takeUforward YouTube channel, popular in India.
Strong Hire … Strong No Hire: The rating scale interviewers use to score a round. The usual steps are Strong No Hire, No Hire, Lean No Hire, Lean Hire, Hire and Strong Hire. See Debrief.
SWESoftware Engineer: The generic job title. Amazon and some others use SDE (Software Development Engineer).

T

Tabulation: Filling a DP table iteratively from the smallest subproblems up. Also called bottom-up DP. See DP.
Tagged: Problems a site lists as reported from a specific company's interviews, as in "Google tagged". Grinding tagged problems is a common last step before an interview.
TCTime Complexity: How running time grows with input size, written in Big O. Often paired with SC in solution posts: "TC O(n), SC O(1)". See SC, TC (Total Compensation), TC (The Candidate).
TCTotal Compensation: Base salary plus bonus plus stock, per year. In offer posts "TC" means this, not time complexity. See TC (Time Complexity), TC (The Candidate), RSU.
TCThe Candidate: Shorthand for the person being interviewed, used in interview feedback and mock-interview notes: "TC asked good clarifying questions." See TC (Time Complexity), TC (Total Compensation).
TC or GTFO: Blind's standing demand that anyone offering an opinion first post their total compensation. See TC (Total Compensation).
Team match: The stage after passing interviews where a candidate talks to teams with open positions before an offer is final.
TILToday I Learned: Introduces something the writer just found out.
TLETime Limit Exceeded: A test case ran longer than the time limit. Usually the algorithm is too slow for the largest inputs, but an infinite loop produces the same verdict. See Big O.
Topo sortTopological sort: An ordering of a directed acyclic graph's nodes where every edge points forward. Used for dependency and scheduling problems. See DAG.
Trie: A tree of characters where each path from the root spells a prefix. Used for prefix lookups and word searches. Pronounced "try".
Two pointers: Two indices moving through an array or string, often from both ends or at different speeds.

U

Upsolve: Solving, after a contest ends, the problems you could not finish during it.
URAUnregretted Attrition: Amazon's term for employees leaving whom the company is content to lose. On Blind it usually refers to reported targets for how many that should be. See Stack ranking, PIP.
USACOUSA Computing Olympiad: A contest for pre-college students with Bronze, Silver, Gold and Platinum divisions.

V

Verbal / written offer: A verbal offer is the recruiter telling you the numbers; the written offer is the signed document. Only the written one is binding.
Virtual contest: A past contest replayed under its original time limit, as practice.
VOVirtual Onsite: An onsite held over video. See Onsite.

W

W / LWin / Loss: A good or bad outcome: "passed the OA, W".
WAWrong Answer: The program ran to completion but its output differed from the expected output on at least one test case.
Weekly / Biweekly: LeetCode's contests: the weekly runs every Sunday, the biweekly every other Saturday. Each has four problems and 90 minutes.
WLBWork-Life Balance: How much of your life the job leaves alone. Team-by-team WLB is one of the most-asked questions on Blind.

Y

YOEYears of Experience: Professional experience, as in "3 YOE".