Airis Labs transforms user generated content into actionable field intelligence for government agencies worldwide.
Our platform empowers frontline analysts with multi modal AI tools that extract clarity from chaos enabling rapid insight across fragmented, unstructured media.
Airis was founded by experienced startup operators and is backed by world class investors.
We are looking for an Applied Researcher, specializing in NLP / Speech / Audio / Computer Vision, to join our TLV based team.
What you'll do
Build and deploy production-grade AI systems around LLMs, VLMs, speech, audio, NLP, and computer vision
Design End to End pipelines- from problem framing and prototyping to evaluation, deployment, and iteration
Work closely with product and engineering to turn complex customer needs into scalable AI solutions
Improve system quality through prompt engineering, evaluations, structured outputs, and model optimization
Own real world AI challenges with meaningful impact
Requirements
Who are you?
3+ years of hands-on experience building and shipping AI/ML systems to production
Strong experience working with modern foundation models (LLMs / VLMs)
Fluent with modern AI tooling- frontier models, prompt design, structured outputs and hands-on experience using Claude Code / Cursor / similar AI agents as an integral part of the development workflow
Strong engineering background with the ability to write clean, maintainable, production quality code that others can easily read, extend, and trust
Experience with cloud and/or on-prem GPU deployments
Strong product intuition and systems thinking
Ownership mentality and ability to operate independently
M.Sc. in CS, EE, Math, Physics, or related field
Nice to have
Video understanding experience
Retrieval/search systems knowledge (vector DBs, embeddings, reranking)
Experience evaluating VLMs and designing benchmarks
Startup background
Familiarity with intelligence or video analytics domains
Why Airis Labs?
Small team of exceptional people
Fast execution and real ownership
Production systems already used by paying customers
Meaningful problems with real world impact
High ownership, healthy work life balance, no hero culture