Separate research roles from product roles
AI and ML job titles blur quickly. A research scientist, applied ML engineer, evals engineer and product engineer using models can all say AI, but the interview loop and daily work differ.
Read the responsibilities before the title. If the role mentions experiments, papers and benchmarks, expect research depth. If it mentions APIs, customers and latency, expect product delivery.
Look for current technical signals
A real listing names the stack, model surface, data type or deployment constraints. Vague phrases like work on cutting-edge AI are not enough.
Company blogs, changelogs, GitHub activity, recent launches and hackathon sponsorships can reveal whether a team is actually building or only hiring for a narrative.
Use projects as proof
For junior and switching candidates, a shipped project often explains skill faster than a credential. A small retrieval app with evals beats a broad portfolio with no working links.
Tie each project to an outcome: reduced review time, improved search quality, lower inference cost or a workflow that someone can use today.
Apply where the role is fresh
AI hiring moves quickly. Listings that are months old without updates may be paused, filled or kept online for pipeline collection.
Prioritize current roles with compensation, location and mode listed. Then use a short note that maps your work to the exact problem in the post.
< read by a human · updated as things change >
browse hackathons