An open research programme for PhD researchers, postgrads, and final-year students. Work on production-grade AI problems with YOTTIX engineers — not toy datasets.
Most academic AI research never leaves the notebook. YOTTIX Labs bridges that gap. You bring the curiosity and rigour. We bring real data, real infrastructure, and real deployment targets. Together, we ship research that matters.
Active problems we're working on. Pick a track or propose your own.
Improving factual grounding in domain-tuned LLMs. Can we reduce hallucination rates below 2% in regulated industries like pharma and finance?
Real-time anomaly detection on resource-constrained hardware. How small can we make a model that still hits 95%+ accuracy on NVIDIA Jetson?
Multi-agent systems that don't hallucinate or loop. Building guardrails, approval gates, and structured reasoning for production agent deployments.
Contrastive learning for alert correlation at scale. Can we build systems that understand incident context better than on-call engineers?
Zero-trust retrieval-augmented generation. Researching chunking strategies, hybrid search, and deployment patterns for air-gapped environments.
Have a research direction that excites you and aligns with our work? Pitch it. We're open to funding and mentoring novel ideas that push boundaries.
Submit your background, interests, and which track excites you. We review weekly.
We pair you with a YOTTIX engineer and scope a 3-6 month research project on real problems.
Access our GPU infrastructure, production datasets, and engineering support. Weekly syncs and code reviews.
Co-author a paper, open-source the work, or deploy it in production. Outstanding researchers get full-time offers.
Who we're looking for:
PhD researchers working on NLP, computer vision, ML systems, or agentic AI
Postgraduate students (M.Tech / MS) in AI/ML with hands-on project experience
Final-year undergrads with published work or strong open-source contributions
Independent researchers with demonstrable expertise in any of our research tracks
Duration: 3-6 months. Remote-friendly. Stipend provided for selected candidates.
Not required, but helpful. We value hands-on building over paper count. A strong GitHub profile or project demo works too.
Selected candidates receive a stipend. The exact amount depends on the track and duration. Outstanding researchers get full-time conversion offers.
Yes. YOTTIX Labs is remote-friendly. We do weekly video syncs and async collaboration through GitHub and Slack.
Typically 20-30 hours/week over 3-6 months. We're flexible around academic schedules and exam periods.
Absolutely. We encourage co-authored publications. Proprietary client data is anonymised, but the methods and findings are yours to share.
Top performers get full-time offers. Everyone leaves with production experience, a published body of work, and a strong reference from our team.
Have questions about the programme?
Ask Yoti →