💸 Earn 20% cashback for every friend you refer who subscribes — Refer & Earn →
Community

Interview & compensation forum

Two things in one place: real interview experiences and compensation packagesfrom ML loops at frontier labs — shared by people who’ve been through them. Sensitive numbers are members-only.

Share your experience
?
loop_survivor
OpenAI · 10h ago
EXPERIENCE

OpenAI onsite — passed, that "infection" coding round

CodingOnsitePass

Made it through the OpenAI onsite and got the pass 🎉 Phone screen was one coding + one design, then onsite was a big coding problem plus a design round — no team-match step at the end. Mostly grinded hack2hire to prep. Sharing exactly what they asked inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
cache_miss
Datadog · PhD · 15h ago
EXPERIENCE

Datadog DL round — implement Focal Loss by hand (fail)

MLDLFail

ML-Eng PhD loop, DL coding round. Asked about Focal Loss and to implement it by hand (details paywalled). Hard, and I didn't pass. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
quant_curious
Amazon · SDE3 · 16h ago
EXPERIENCE

Amazon SDE3 onsite — LC445 variant + a vending-machine LLD (fail)

CodingLLDFail

SDE3 loop, didn't make it. Phone was an add-two-numbers variant; the low-level design was a fun one about auto-assigning the nearest stocked machine. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
grind_ghost
Citadel · 17h ago
EXPERIENCE

Citadel GQS phone screen — the whiteboard "find all K" twist

CodingPhoneQuant

All-whiteboard chat with a quant dev, no IDE. Felt less about the code and more about whether one API is expressive enough. A headhunter reached out to me for this one. The exact problem + the constraint that tripped me up is inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
null_ptr
Abnormal AI · 22h ago
EXPERIENCE

Abnormal AI screen — code-review + Decorator pattern on a security codebase

CodingDesignFail

Hardest screen I've had. A security-event-processing codebase (collection, ranking, rule-based threat leveling, alerts, APIs, dataviz): code review + design patterns (Decorator) + implement. Failed. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
ml_maybe
DoorDash · 23h ago
EXPERIENCE

DoorDash process — HR + an algo/fundamentals + behavioral R1

BehavioralMLProcess

Sharing the process shape more than a single question. HR call first, then R1 mixed algorithm/model fundamentals with behavioral. Structure your answers conclusion → details → summary and quantify your project impact. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
anon_grad
Google · 1d ago
EXPERIENCE

Google onsite — the "piano finger movements" sliding window

CodingOnsite

Onsite coding, and I didn't fully finish the follow-up. Wondering if 2 hire + 1 lean-hire + 1 hire/strong-hire is still enough for an offer. The exact question inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
stack_trace
Electronic Arts · SDE2 · 1d ago
EXPERIENCE

EA SDE2 — OA then an AWS + resume video interview (fail)

CodingOAFail

SDE2. OA had 2 DSA questions (paywalled); the video interview was AWS services + resume experience. Didn't pass. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
segfault_sam
Anthropic · 1d ago
EXPERIENCE

Anthropic onsite — "why Anthropic" + a SQL/data round

SQLOnsiteData

A non-SWE-flavored loop — lots of culture fit plus a data-analysis round that needed some cloud-capacity intuition, and some SQL. Commenters were surprised there was no coding/SD round. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
offer_hunter
ByteDance / TikTok · 1d ago
EXPERIENCE

TikTok USDS phone — number of islands (fail)

CodingPhoneFail

Technical phone for the USDS org, and honestly I didn't vibe with it — never even caught the interviewers' names. Ended up failing. The (very standard) question is inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
edge_caser
Millennium · New grad · 1d ago
EXPERIENCE

Millennium QR-ML phone — a modeling problem, probability heavy

MLQuantPhone

New-grad QR-ML. 40-min phone with a quant researcher, one modeling problem (paywalled). Tips: multi-step logic, practice modeling, prep pseudo-code, expect probability. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
base_case
Anthropic · 1d ago
EXPERIENCE

Anthropic "Coding Interview with Claude" — the new AI-tools format

AI-assistedOnsiteInfra

Got Anthropic's newer format where the whole point is writing/reviewing code with AI tools (tool access provided, no setup needed). Applied online; likely an infra role. Sharing the exact prompt they gave inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
kv_kid
Rippling · 1d ago
EXPERIENCE

Rippling phone — "does this transaction break a rule?" (AI-assisted by default)

CodingPhoneAI-assisted

Phone-screen coding that was less about the algorithm and more about explaining your design and walking examples. Interesting bit: the interview was AI-assisted by default — they didn't even ask, it was just on. Details inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
red_flag_ray
OpenAI · L? · 1d ago
EXPERIENCE

OpenAI phone — versioned KV store + social graph (pass)

System DesignPhonePass

Easy-rated 60-min phone and I passed 🙌 One main problem plus a follow-up, design-friendly, positive vibe from the interviewer. Someone asked SDE vs MLE — this was the general track. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
dp_dave
Shopify · 1d ago
EXPERIENCE

Shopify phone (referral) — a "rover" problem, then scale it to 1M

CodingPhoneScale

Referral phone screen: a rover problem close to common practice ones, with a follow-up to scale from a few rovers to 1M+. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
recruiter_dodger
Blink Health · Senior · 1d ago
EXPERIENCE

Blink Health onsite — OA with a prompt-engineering bonus round

System DesignOnsiteOA

Senior SWE loop. The OA was a grab-bag (MCQs across SWE + AI, a DSA one, an API problem, and an optional prompt-engineering one). Round 2 had a design + a deep dive on past project choices. Full breakdown inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
vo_veteran
Rippling · Staff+ · 2d ago
EXPERIENCE

Rippling Staff+ phone — traditional vs LLM design, your call (pass)

System DesignPhonePass

Passed! 75-min Staff+ design where they let you pick traditional vs an LLM approach. Got a latency-estimation curveball (estimate in ~20s). Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
heap_hazel
Databricks · 2d ago
EXPERIENCE

Databricks phone — solved it but a dumb bug cost me the follow-up (fail)

CodingPhoneFail

Rejected 3 business days later. I actually got a one-pass solution, but a simple bug ate my time so I never reached the follow-up — probably why. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
late_night_lc
Snowflake · IC4 · 2d ago
EXPERIENCE

Snowflake phone — coding + system design back to back (pass)

CodingSystem DesignPass

Passed! Two rounds same day, back to back: coding then system design. Trying to figure out what IC4 here maps to at Google/Meta. What I can share of the coding round is inside (rest was paywalled).

Full writeup locked
See the exact questions, answers and follow-ups.
?
ghosted_again
Intuit · Staff+ · 2d ago
EXPERIENCE

Intuit Staff+ phone — pick traditional vs LLM, sales-data stream (pass)

System DesignPhonePass

Passed after a scare. 75-min Staff+ design: traditional vs LLM (I went LLM), plus a sales-data-stream problem. Interviewer screenshotted my design; I thought I'd bombed but got the yes 4 days later. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
binary_bri
Roblox · 2d ago
EXPERIENCE

Roblox MLE phone — dependency build order (fail)

CodingPhoneMLE

MLE phone screen, failed. Got a lower-frequency build-order problem; writing it out here since older posts were unclear. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
sim_solver
Pinterest · 2d ago
EXPERIENCE

Pinterest phone — an unusual "several arrays" problem (fail)

CodingPhoneFail

Technical phone, failed. Got what looked like a fresh, unseen question — people in the comments debated whether it even needs an algorithm or is just careful iteration. Also had an ML-basics question. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
package_pete
Retell AI · 2d ago
EXPERIENCE

Retell AI (startup) — design a SQL-parsing system, boundaries game

System DesignStartupML

AI-voice startup (Chinese founder, several Chinese engineers). R1 was system design: a system that processes various SQL syntax, and they keep pushing for more features so you're constantly setting scope boundaries. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
loop_survivor
Jane Street · 2d ago
EXPERIENCE

OpenAI onsite — passed, that "infection" coding round

CodingPhoneFail

Technical phone, failed. The tree problem itself was paywalled on the source; noting it here for the record.

Full writeup locked
See the exact questions, answers and follow-ups.
?
panel_prey
Fuse Energy · 2d ago
EXPERIENCE

Fuse Energy onsite — the linked-list cycle with a void twist (fail)

CodingOnsiteFail

Onsite, and I got rejected even after the interviewer literally said "I'll buy you this design" mid-round — still not sure what they wanted. My advice: actually write the thing out, it's harder than it looks. Exact setup inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
quant_curious
Amazon · PhD · 2d ago
EXPERIENCE

Amazon Applied Scientist — science depth + a manager-tree coding (fail)

MLCodingFail

CS PhD, 5 YOE, DS/MLE background — two OA-style rounds and a 1-year HR freeze after 😮‍💨 R1 was science depth (PEFT, vLLM inference, image-text models); R2 a coding problem I ran out of time on. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
grind_ghost
Microsoft · 3d ago
EXPERIENCE

Microsoft phone (referral) — HLD+LLD URL shortener, find the bugs

System DesignPhone

Referral technical phone. HLD + LLD of a URL shortener plus a find-the-bugs task (details paywalled). Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
onsite_ollie
Anchorage Digital · 3d ago
EXPERIENCE

Anchorage Digital onsite — OOD + a key-vault design (fail)

System DesignOnsiteFail

Took this one mostly for practice and still failed lol. Coding was a pretty normal OOD; the design was the unusual part. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
ml_maybe
Salesforce · MTS · 3d ago
EXPERIENCE

Salesforce MTS onsite — Course Schedule II + a prime-split DP (pass)

CodingSystem DesignPass

Passed the MTS loop — hardest-rated but doable. DSA round had LC210 plus a prime-split counting problem; SD round was a parking/car-rental booking system. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
anon_grad
Otter AI · 3d ago
EXPERIENCE

Otter AI — design a KV cache in Java, ran out of time on transactions

CodingDesignPhD

PhD-track loop. Designed a KV cache in Java; debated whether the value can be int or string (I went Object wrapper), then ran out of time on the transactions part. Sardonic note: Python would've been comfier. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
cache_miss
Etsy · 3d ago
EXPERIENCE

Etsy "recruiter call" that was secretly a tech screen

MLRecruiterScreen

Booked what I thought was a normal recruiter chat and instead got grilled on ML with zero warm-up. Felt like AI on both ends — buzzwords on your resume can genuinely backfire now. The questions inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
segfault_sam
Meta · 3d ago
EXPERIENCE

Meta AI-assisted coding — bug-fix → implement → optimize (fail)

CodingAI-assistedFail

Meta's newer AI-assisted coding format and it caught me off guard. Looks like one problem but it's got hidden follow-ups you must finish in time — very different from the traditional "we'll tell you how many questions + take a break" flow. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
offer_hunter
SIG · 3.5 YOE · 3d ago
EXPERIENCE

SIG onsite — implement std::vector from scratch

CodingOnsiteC++

3.5 YOE, C++ role. Phone was implementing std::vector (placement new, operator new, resize, move) — finished in 45 of 90 min, chatted the rest. Been searching 4 months, lots of finals, no offers yet 😮‍💨 Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
null_ptr
Amazon · 3d ago
EXPERIENCE

Amazon in-person onsite — actually pleasant (LP + OOD)

OODOnsiteLP

In-person onsite and honestly it was nice — food, coffee, good vibes (no dreaded "dog" round). Someone asked what "Pizza" meant in my notes — it's just OOD. Structure inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
kv_kid
Meta · PhD · 3d ago
EXPERIENCE

Meta onsite — 2 research design rounds + behavioral

MLResearchOnsite

MLE/Research PhD loop: two research-design rounds (scenarios + language-model topics, incl. a linear-attention discussion) and one behavioral. Recruiter seemed dedicated to MSL/MRS. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
stack_trace
Uber · 3d ago
EXPERIENCE

Uber phone — two mediums, nailed the complexity Qs (pass)

CodingPhonePass

Passed the technical phone. Two medium LC-style problems and I got all the complexity follow-ups right. People are asking if Uber is frozen (loops finishing with no team). Details inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
recruiter_dodger
Lyft · 3-4 YOE · 4d ago
EXPERIENCE

Lyft phone — 50 min, all tests passed (pass)

CodingPhonePass

Passed! ~50 min, mostly problem-solving, all test cases green, next-steps email within 30 min. Likely a junior req (VO had laptop + SD + HM). Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
edge_caser
Stripe · L2 · 4d ago
EXPERIENCE

Stripe L2 onsite — 4 rounds, references checked, still a no

CodingOnsiteFail

Full L2 loop and it stung: strong-ish scores across debug/integration, but the AI-coding round I only finished 1 question, got ghosted on an HR call, references (2 managers) called and deemed "insufficient signal," rejected next day as a final HC call. The rounds + the red-flag debate inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
late_night_lc
Optiver · Intern · 4d ago
EXPERIENCE

Optiver OA (intern) — OOD + a reaction test + a Hanoi-like game

OODOAIntern

Intern OA, 75 min. One OOD-ish problem (write class __init__ + a method), a recognition/reaction test (match the identical option from 4 in ~1s), and a Hanoi-like game. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
red_flag_ray
Google · 4d ago
EXPERIENCE

Google behavioral round — old-school "tell me a time" (worried)

BehavioralOnsite

The BQ round with a veteran interviewer (IIT, 10+ yrs at Google), 3-4 "tell me a time" questions, ended early. I'm worried this is my weak link since the rest felt solid. Comments say BQ rarely sinks people. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
sim_solver
Akuna Capital · Intern · 4d ago
EXPERIENCE

Akuna OA (intern) — a stock-price state machine

CodingOAIntern

Intern OA, 120 min, one question. Stock-price records → apply trading limits → track state / best profits. First figure out the data format and constraints, then state-machine or traversal. Moderate. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
vo_veteran
Roblox · Senior · 4d ago
EXPERIENCE

Roblox phone — a 30-min HM chat first, no system design

HMPhoneSenior

Senior-level, and interestingly there was no system design round. R1 was a 30-min hiring-manager chat (intro + a project). Now waiting after the screen and wondering how long that usually takes. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
panel_prey
Akuna Capital · Intern · 5d ago
EXPERIENCE

Akuna VO (intern) — behavioral + Python + on-the-spot code mods

BehavioralVOPython

Intern VO. STAR behavioral, Python pros/cons, walked through my OA code, and fixed-point money issues (Decimal vs float). Follow-ups make you modify code live. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
ghosted_again
Zoox · 5d ago
EXPERIENCE

Zoox phone — C++ trivia to start

CodingPhoneC++

Phone screen kicked off with C++ trivia before the coding (rest was paywalled). The bit I can share is inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
onsite_ollie
Trexquant · 5d ago
EXPERIENCE

Trexquant — friendly chat, LC295 median of stream (fail)

CodingPhoneFail

Full-time via headhunter. Friendly Chinese interviewer, detailed resume questions, then LC295 (Find Median from Data Stream) and a map vs unordered_map chat. Didn't pass. Notes on sponsorship inside.

Full writeup locked
See the exact questions, answers and follow-ups.
?
package_pete
Google · Intern · 5d ago
EXPERIENCE

Google intern VO — two coding rounds back to back

CodingOnsiteIntern

Intern virtual onsite: two coding rounds straight, no system design. Someone asked what "back to back" means — it just means two straight hours, nothing special. Inside.

Full writeup locked
See the exact questions, answers and follow-ups.