Becoming a Research Engineer at a Big LLM Lab
5 hours ago
- The author secured a research engineer role at Mistral after 18 months of strategic career development, emphasizing the difference between tactical (low-effort, high-return) and strategic (high-effort, compounding) actions.
- Personal timeline: defined goals, networked, upskilled (LeetCode, CS fundamentals), failed first application, quit job, joined Recurse Center, learned Rust, contributed to open source, wrote a paper, then applied in batches and got offers.
- Goal setting is crucial: define clear criteria for the ideal role (e.g., career inflection, rare skills, ownership, senior peers, IC path) to focus efforts and say no to compromises.
- Application playbook: batch applications, use network for referrals and insider info, schedule lower-stakes interviews first to gain practice.
- Networks rely on weak ties: build genuine friendships by doing interesting things, sharing them, and being open; online (Twitter) and offline (cohort programs) are effective.
- Career momentum requires building rare and valuable skills (career capital) and demonstrating an upward trajectory through projects, OSS contributions, publications, and CV entries.
- Interview types include initial screen, takehome, online assessment, coding interviews (LeetCode, API/system design, real-world), system design, culture fit, quiz, hiring manager, and reference checks—each with specific preparation strategies.
- The mental game involves managing rejection, uncertainty, and stigma; staying motivated by remembering the leverage and possibilities of job seeking.
- Tactics boil down to gaining an information advantage and acting on it, while strategy focuses on long-term skill building and proof points.