Refery blog
AI hiring
Hiring AI engineers, forward-deployed engineers, and applied ML teams.
Recruiters and Platforms That Specialize in AI Startups (2026)
Six recruiters and platforms that publicly focus on AI startup hiring, what each one says it does, what it publishes about fees, and how to brief an AI role so any of them can work it. Checked September 19, 2026.
AI infrastructure or AI application engineer: define the bottleneck
Separate AI application delivery from shared infrastructure work by tracing the actual bottleneck in the product.
How to run a coding interview that allows AI tools
Set clear AI-tool rules and assess an engineer's understanding, verification, debugging, and judgment during a coding exercise.
More in ai hiring
12 guidesHow to assess inference engineering candidates
Assess inference engineers through workload characterization, latency and throughput tradeoffs, controlled experiments, and production judgment.
Reviewing AI-assisted recruiting recommendations
Review recruiting recommendations the way Refery's search examples demand: distinguish work evidence, fit judgments, candidate interest and unresolved interview questions.
An interview exercise for engineers building enterprise AI agents
Assess enterprise agent engineers using a fictional workflow with tool permissions, ambiguous outcomes, recovery, and user oversight.
Defining a head of AI role across data, research, and product
Set a head of AI mandate by defining decisions across product, data, research, infrastructure, and evaluation.
How to define retrieval engineering work in an AI product
Scope retrieval engineering across document ingestion, relevance, access control, freshness, and evaluation for an AI product.
Machine learning engineer or data scientist: scope the next outcome
Choose a machine learning hire by separating problem formulation, model experimentation, production integration, and ongoing operation.
How to assess a research engineer's experimental work
Review research engineering evidence through experiment implementation, reproducibility, comparison quality, and honest interpretation.
Research scientist or AI engineer: which does your startup need?
Choose an AI hiring mandate by separating research uncertainty from product integration, evaluation, and production ownership.
How to scope reinforcement-learning environment engineering
Define reinforcement-learning environment engineering through task dynamics, observations, actions, rewards, termination, and evaluation integrity.