Interview Experiences

Technical screen (60 min): Networking fundamentals + one coding problem about parsing HTTP headers efficiently. Onsite (4 rounds): 1. Low-level systems — questions about TCP/IP, TLS, DNS, BGP routing. Very deep. 2. Coding — implement a simplified HTTP request router with wildcard matching 3. System design — design Cloudflare's DDoS mitigation layer 4. Engineering culture — how Cloudflare operates at the edge, Rust adoption, mission

Recruiter screen (30 min): Notion's mission, product passion, background. Technical screen (60 min): Build a simplified version of Notion's block-based document system — blocks have types (text, heading, todo) and can be nested. Live coding in a shared editor. Product screen (45 min): How would you improve Notion's collaboration features? Think like a PM. Onsite (3 rounds): 1. Technical deep dive — extend the block system with real-time sync support 2. System design — design Notion's database feature at scale 3. Hiring manager — values, career path, how you approach ambiguity

Applied via referral from a current employee. Recruiter screen (30 min): Role overview, remote work discussion. GitHub is fully remote. Technical screen (60 min): Two problems — one about implementing a simplified git diff algorithm (line-by-line), one about graph connectivity (finding connected components). Did well on both. Onsite loop completed last week — results still pending.

Recruiter screen (30 min): Role overview and background. Technical screen (60 min): One coding problem — design a metrics aggregation system that supports counters, gauges, and histograms. Very relevant to what Datadog actually builds. Onsite (4 rounds): 1. Coding — implement a log parser that extracts structured data from raw log lines 2. System design — design a distributed tracing system (basically design Datadog APM) 3. Debugging — given a live (simulated) production incident with metrics graphs, diagnose and describe a remediation plan 4. Behavioral — past technical decisions, cross-team collaboration

Recruiter screen (30 min): Background, cleared security clearance potential (some roles require it). Karat interview (60 min): Third-party coding interview — two medium problems, timed. Felt impersonal. Onsite (decomposition + coding): 1. Decomposition — given a vague problem (design a hospital scheduling system), break it down into components and APIs. 60 min, open-ended. 2. Coding — implement a simplified version of one component from the decomposition exercise. 60 min. 3. Culture — Palantir's mission, working in high-stakes environments, ethics in data use. Rejected after coding round. Feedback: "strong decomposition but implementation didn't match design."

Technical screen (90 min): Pair programming in a real editor (VSCode) — build a simplified e-commerce cart system with discount rules. Live collaboration, not whiteboard. Values screen (45 min): Shopify's "Impact" values, what makes great engineering, how you work remotely. Onsite (3 rounds): 1. System design — design Shopify's checkout flow at Black Friday scale 2. Coding — extend the cart system from the technical screen with new requirements 3. Hiring manager + career discussion — growth plans, interests, how you handle ownership

Recruiter screen (30 min): Role overview + background check on crypto experience (you don't need it but helps). Technical take-home (3–4 hours): Build a simple blockchain transaction validator in the language of your choice. They care about code quality and test coverage, not just working code. Technical review (60 min): Walk through your take-home, defend your design decisions, then one live coding problem (implement a simple LRU cache). Values interview (45 min): Why crypto, mission alignment, how you handle uncertainty.

Recruiter email → technical phone screen scheduled. Technical phone screen (60 min): Two problems — one easy string manipulation, one medium graph (find the shortest path in a city map). Did well on both. Recruiter said "we'll be in touch within a week about next steps." Never heard back. Followed up twice over 3 weeks. No response.

Recruiter screen (45 min): Very detailed discussion about career goals and leadership experience. Netflix is selective about culture fit from the very start. Hiring manager screen (60 min): Half technical discussion about past architecture decisions, half about how I handle "freedom and responsibility." Technical screen (60 min): System design only — design Netflix's video encoding pipeline for a new device category. No coding at this stage. Onsite (5 rounds): 1. System design — design Netflix's recommendation system at 200M users 2. System design — design the CDN layer for video delivery 3. Coding — two medium problems (rate limiting, distributed counter) 4. Values — intense discussion around past decisions, handling conflict, ownership 5. Leadership — how have you influenced engineers outside your team?

Phone screen (45 min): One coding problem — design a simple booking conflict detector (given a list of reservations, find overlapping ones efficiently). Onsite (6 rounds — yes, six): 1. Coding — implement a simplified version of React's reconciliation diff algorithm 2. Coding — design an autocomplete system, then optimize for latency 3. System design — design Airbnb's search feature with geo-filtering and ranking 4. Cross-functional — how do you work with product managers and designers? 5. Estimation — how would you measure the success of a new feature? 6. Experience — deep dive into most complex past project Rejected after onsite. No specific feedback given.

Recruiter screen (30 min): Background and role alignment. Phone screen (60 min): One coding problem — LRU cache implementation. Straightforward. Onsite loop (4 rounds × 60 min): 1. Coding — design and implement a file system with mkdir, ls, and cat commands 2. Coding — graph problem (course schedule / topological sort) 3. System design — design a distributed key-value store 4. Behavioral — tell me about yourself, past projects, collaboration with difficult people Overall more relaxed than Google/Meta. Interviewers are approachable and give hints if you're stuck.

Recruiter chat (30 min): Role overview, compensation discussion (Stripe is very transparent about pay upfront). Technical screen (60 min): Pair programming in a shared editor — build a simplified in-memory rate limiter supporting multiple algorithms. Not a LeetCode problem — more of a practical design + coding exercise. You're expected to write working code, not pseudocode. Onsite (4 rounds): 1. Systems — design Stripe's payment processing pipeline, focus on reliability and idempotency 2. Coding — two practical problems (CSV parser, then extend it with filtering/aggregation) 3. Integration — design Stripe's webhook delivery system with retries and ordering guarantees 4. Values — culture fit, what kind of engineer are you, how do you handle ambiguity

OA (2 hours): Two coding problems on HackerRank + a Work Style Assessment (personality questionnaire). Problems were medium difficulty — one array manipulation, one graph/BFS. Virtual loop (5 rounds × 60 min — bar raiser included): 1. Coding + LP — two problems (two sum variant, string parsing) + 2 leadership principle questions 2. Coding + LP — binary search + tree traversal + 2 LP questions 3. System design — design Amazon's product recommendation engine 4. LP focus (bar raiser) — deep dive into a past project, lots of "why" follow-ups 5. Coding + LP — DP problem (coin change variant) + LP questions Amazon is 50% LP. Every round has at least two behavioral questions.

Recruiter screen (30 min): Background, motivation, timeline. Very friendly. Technical phone screen (60 min): One coding problem — serialize and deserialize a binary tree. Got it right but was slow on the follow-up (could you do it with less memory?). Onsite loop (4 rounds): 1. Coding — two problems: valid parentheses variant + merge k sorted lists 2. Coding — design a data structure supporting insert, delete, getRandom in O(1) 3. System design — design Facebook's news feed 4. Behavioral — past projects, disagreements with managers, biggest failure Result: rejected at the coding round. Feedback was "strong in design but coding speed needs improvement."

Phone screen (45 min): Two medium-difficulty LeetCode-style problems — sliding window and binary search on a sorted matrix. Moved on the same day. Virtual onsite (5 rounds × 45 min): 1. Coding — graph BFS/DFS, find all paths between two nodes with constraints 2. Coding — DP, maximum profit with cooldown (LeetCode 309 variant) 3. System design — design a URL shortener at Google scale (billions of URLs) 4. Behavioral — leadership and conflict resolution using STAR format 5. Googleyness — values and culture fit, why Google, what motivates you Received verbal offer 5 business days after onsite.