#036 ML Борис Шарчилев. Как пройти собеседование в ML-команду

#036 ML Борис Шарчилев. Как пройти собеседование в ML-команду

#036 ML Boris Sharchilev. How to pass an interview for an ML team

by Machine Learning Podcast · View the channel on PSI

Humanities, Social Sciences & Thought AI — Models & Research UYQArtificial intelligence (AI)UYQMMachine learning
🎙 Machine Learning Podcast 👥 5K 📅 December 28, 2021 ⏱ 51 min 👁 9K 📄 expert opinion 🧭 2026-10-09
Available in: English (current) Français

Keywords

ML interviewhiringdata scienceengineeringcareer advice

Summary

In this podcast episode, host interviews Boris Sharchilev, head of ML at Yandex Fintech, about how to prepare for and pass ML team interviews. Boris shares his experience from conducting over 100 interviews and discusses the typical hiring pipeline, from initial recruitment to technical sections and final interviews. He distinguishes between data scientists and ML engineers, emphasizing that engineering skills are crucial. The episode covers the structure of ML interviews, including algorithmic coding tests and practical case discussions. Boris reveals that he prefers discussing real-world ML problems over theoretical textbook questions, as this better reflects the actual job. He also addresses common concerns, such as the need to know specific tools, and explains that adaptability matters more than familiarity with proprietary systems. The conversation includes tips on preparing for algorithm challenges using platforms like Codeforces and LeetCode, and highlights that soft skills are often evaluated during the final interview. Overall, the podcast provides insider perspectives on what employers like Yandex look for in ML candidates and how to navigate the hiring process effectively.

173 words

Critical Evaluation

Value of the Information & Strength of the Argument

The value of the information lies in its practical, first-hand nature. Boris offers concrete insights into the interview process at a major tech company, such as the importance of being able to reason through real ML cases rather than recite theoretical knowledge. He argues persuasively that coding and engineering proficiency are essential, and that the ability to learn and adapt to new tools is more valuable than knowing specific frameworks. His argumentation is coherent and grounded in personal experience, though it remains anecdotal and lacks statistical evidence. He also debunks the idea that there is a fixed checklist of must-have skills, instead emphasizing adaptability and problem-solving. The discussion is solid, with clear reasoning, and effectively conveys the nuances of ML hiring.

Scientific Rigor, Source Quality, Title Accuracy

The video does not present a scientific study but rather draws on the author’s professional experience as a hiring manager. No external sources, such as articles or studies, are cited within the episode itself. The description provides links to practice platforms (Codeforces, HackerRank, LeetCode) and the podcast’s distribution channels, which serve as references for interview preparation. The title accurately mirrors the content: it promises advice on passing ML interviews, which the guest delivers. There is no mismatch between title and content. However, the scientific rigor is limited because the advice is based on individual impressions and does not reference broader research or data.

239 words

Title / Content Match

The title accurately reflects the content, which focuses on the ML interview process and preparation.

Quality & Reliability

7/10

The guest, a hiring manager with over 100 interviews conducted, offers practical, experience-based advice. The content is credible but subjective and lacks empirical validation.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

Contribution & Novelties

The video provides a rare insider perspective on ML hiring at a major Russian tech company, offering practical advice on how to prepare for technical interviews, what distinguishes data scientists from ML engineers, and how to approach case-based problems. It debunks common myths about the need for specific tools and emphasizes adaptability. The originality lies in the guest’s extensive interview experience and his preference for realistic scenarios over theoretical questions.

Pour aller plus loin :

  • System design interview — Useful for senior ML roles, as the video mentions architectural interviews.
  • MLOps — Relevant to the production ML engineering aspects discussed.
  • Kaggle competitions — A popular platform for honing practical ML skills, similar to Codeforces/LeetCode but for ML problems.

118 words

Radar Profile

The radar profile shows high scores in information quantity and quality, moderate technical depth, and a reasonable reliability score, reflecting a balanced discussion with practical value.

Reliability 7/10