Get Started
Hunting for you, 24x7Get the iPhone app

Your hunt runs in your pocket, and tells you the moment a strong match opens.

Download on theApp Store
Anuvaya LabsPosted 8 days ago

Member of Technical Staff - Applied AI Research

Who we'd reach at Anuvaya LabsFound and reached by Power Apply
Best leadHiring manager

Leads the team this role joins

Talent partner

Hires for this team

Senior teammate

On the team, can refer you

Power Apply

From your own Gmail and LinkedIn.

About the role

This is a full-time, on-site applied AI research role in New Delhi for an engineer who fine-tunes language models for Hindi and other Indian languages and makes them reliable in production. It suits people who can train and evaluate models, and who can also build the prompt, context and retrieval systems that serve real users. The work centers on Hindi, Hinglish and other Indian languages, where current foundation models fall short. The team is small and works from a hacker house in Vasant Kunj, and the role is on-site only with no remote or hybrid option.

What you’ll do

  • Fine-tune models to improve fluency, accuracy and domain understanding in Hindi, Hinglish and other Indian languages
  • Build fine-tuning pipelines covering data curation, training runs, evaluation and deployment
  • Design prompt, context and multi-turn conversation structures with the team
  • Design retrieval strategies that deliver the right domain data at the right time
  • Run experiments on how fine-tuning, context layout, prompt wording and tool design affect output across languages
  • Work with the evaluation side to measure quality, especially subjective, language-dependent quality
  • Track frontier model developments and adopt new features quickly

What they’re looking for

  • Experience fine-tuning LLMs, including managing datasets and evaluating results
  • Understanding of Indic NLP issues such as tokenization, code-switching, script variety and model failure modes
  • Experience running LLMs in production with prompt engineering, context management and retrieval-augmented generation
  • A practical sense of the gap between a demo and a production system, including context windows, hallucinations and inconsistent behavior
  • Strong engineering skills and a record of shipping code
  • Experimental rigor: designing experiments, controlling variables and judging significance

Nice to have

  • Fine-tuning on Indic or multilingual data
  • Deep experience with Claude, Gemini or other frontier model APIs
  • Familiarity with Elixir, TypeScript with Bun, PostgreSQL and NATS
  • Published work or substantial projects in applied NLP, multilingual models or knowledge-grounded generation
  • Having read the team's whitepapers on the Realtime Context Engine and context splitting, with views on what they got wrong

What’s on offer

  • On-site only, at a hacker house in Vasant Kunj, New Delhi
  • Small, flat team where engineering, design and product work closely together
  • Heavy focus on building in public through open-source work and detailed blog posts
  • The role is not a standard nine-to-five, and after-hours thinking is expected

Skills

Questions about this role

Is this job remote?

No. The role is on-site only, with no remote or hybrid option.

What languages does the work cover?

Hindi, Hinglish and other Indian languages.

Do I need prior fine-tuning experience?

Yes. Experience fine-tuning LLMs, including managing datasets and evaluating results, is listed as a requirement.

Is the job full-time?

Yes.