← Back to Directory

Cohere

Mid-sized

Screna has 12 Cohere interview write-ups. The most reported round is the technical phone screen (8).

12 interview notes · updated Jun 2026

Cohere·Machine Learning Engineer·Onsite - Multi Round

Jun 2026
Round 3 at Cohere for an ML Engineer role involved presenting a research paper, and they pushed well beyond surface-level understanding into the paper's weaknesses and how the field had moved on since publication. Pretty intense if you're used to just summarizing results.
  • What are the limitations of the experimental setup in the paper you presented?
  • How has the field progressed since this paper was published, and what would you change or extend given those developments?
  • Walk me through how supervised fine-tuning works and where it tends to fall short.
  • Explain the RLHF pipeline and the tradeoffs involved in reward model design.

“This is where I got a bit tripped up.”

View Post

Cohere·Machine Learning Engineer·Onsite - Coding / Algorithms

Jun 2026
Cohere ML Engineer live coding round where you're handed a messy Python text game and told to rip out the heuristic if/else logic and replace it with something cleaner, all while AI tools are fair game. The meta-layer here is that they're not just watching you code, they want to talk through how you used the AI, where you trusted it, and where you didn't.
  • You're given a terminal-based Python text game with several hundred lines of heuristic if/else logic. Replace the implementation while keeping all externally observable behavior identical, using AI coding tools if you want. Demonstrate equivalence by running end-to-end scenarios.
  • Walk through your prompting strategy for this task. How did you decide what to ask the AI tool to do versus what to handle yourself?
  • How did you verify that the new implementation is actually equivalent to the original? What does equivalence even mean in this context?
  • Where did you trust the LLM's output and where did you feel you needed to intervene manually? What signals told you to step in?
  • How would you harden this kind of AI-assisted refactor if it were going into a production system?

“This was the whole round basically.”

View Post

Cohere·Machine Learning Engineer·Technical Phone Screen

Jun 2026
Cohere ML Engineer interview that was essentially one long open-ended design conversation about how you'd make an LLM better at multi-step reasoning. No leetcode, no trivia, just a sprawling discussion that covered basically every technique in the book from data curation all the way to latency budgets.
  • Say you have an LLM that struggles with tasks requiring multi-step reasoning, like math word problems, agentic tool use, or code generation with planning. Walk through how you'd improve it end to end, covering data, training, inference, evaluation, and cost trade-offs.

“This is one of those questions that sounds like it has a clean answer until you're actually in it.”

View Post

Cohere·Software Engineer·Technical Phone Screen

May 2026
Interviewed for a data engineering role at Cohere. Just one question I can remember clearly, but it was a good one that made me think harder than I expected.
  • If a conversation spans across two days, how would you design your data pipeline to be idempotent?

“The two-day span detail is what trips you up.”

View Post

Cohere·Software Engineer·Technical Phone Screen

May 2026
Interviewed for a data engineering role at Cohere. One technical question on deduplication strategy, felt like a system design probe more than a SQL quiz.
  • How would you approach deduplication across an entire table, not just within a single day's ingestion batch?

“I started with the daily dedup case and then had to mentally shift gears to think at table scale.”

View Post

Cohere·Software Engineer·Hiring Manager Screen

May 2026
45-minute live session with the hiring manager at Cohere for a data engineer role. The whole thing was a walkthrough of actual work, which felt more like a code review than a traditional interview.
  • You're filling missing values with 0 in your pipeline. Why not just drop those records entirely? And why not leave them as NULL rather than replacing them?

“This one got me more than it should have.”

View Post

Cohere·Machine Learning Engineer·Technical Phone Screen

Apr 2026
Cohere ML engineer interview with a coding question on decoding strategies. Pretty implementation-heavy, they gave you a mock GPT class and wanted you to build greedy and top-k decoding from scratch with exact output assertions to pass.
  • Given a simulated GPT-like class with a fixed vocab size, an EOS token, and a deterministic forward method returning logits, implement two autoregressive decoding functions: one using greedy decoding (argmax at each step) and one using top-k sampling (restrict to top-k logits, renormalize, then sample). Both should stop at EOS and return only the generated tokens. Your implementation must pass exact sequence assertions for a given input.

“The greedy part was straightforward, argmax in a loop until you hit EOS, not much to overthink.”

View Post

Cohere·Software Engineer·Technical Phone Screen

Apr 2026
One technical question from what seemed like a data engineering screen at Cohere, focused on Spark internals. Short but not easy.
  • A Spark job is running and one worker node is experiencing data skew. How do you diagnose and fix it?

“This is the kind of question where you think you know the answer and then halfway through your explanation you realize you're describing the problem not the solution.”

View Post

Cohere·Machine Learning Engineer·Technical Phone Screen

Apr 2026
Cohere ML Engineer interview, round 2 was a coding round focused on ML fundamentals. I came in expecting to write PyTorch and got hit with a NumPy implementation question instead. Not my best showing.
  • Implement top-K sampling from scratch using NumPy.

“I froze.”

View Post

Cohere·Machine Learning Engineer·Technical Phone Screen

Apr 2026
Oral technical screen for an MLE role at Cohere covering basically the full ML stack. Four distinct topic areas, each with follow-up questions that go pretty deep. The RL/post-training section gets the most weight, so don't treat it as a quick review.
  • Walk me through how you approach data curation and mixing decisions when pre-training a large language model.
  • How does tokenization choice affect model training and downstream performance?
  • Compare DPO, PPO, and GRPO for post-training alignment. When would you pick one over another?
  • What are the alignment trade-offs you're making when you apply RLHF, and how do you think about reward hacking?
  • Explain how you'd set up distributed training across data parallelism, tensor parallelism, pipeline parallelism, and FSDP. What are the tradeoffs between these strategies?
  • How do AdamW and Lion differ, and when might you prefer one over the other for LLM training?
  • How do you design evaluations that are reliable and not contaminated, especially for capability vs alignment distinctions?

“This is where I felt least prepared.”

View Post