About Intellimation.ai Intellimation.ai builds AI that runs in production inside banks, asset managers and hedge funds. Founded in 2018 and profitable, we have 40+ autonomous workflows live at 30+ Tier-1 financial institutions across reconciliation, margin and collateral, structured products, CLO and credit, ISDA and prime brokerage agreements, and client onboarding. We are not a model lab and we do not sell a chatbot. We work at the inference layer: a neuro-symbolic reasoning and extraction engine, built on financial-domain knowledge, that sits over any frontier or open-weight model and turns it into an auditable, governed system a regulated firm will sign off. We specialise models to each workflow, so clients get frontier-level accuracy on infrastructure they control. Our line is simple: proof, not pilots. If you want your research to end up in front of a bank's model-risk committee rather than in a demo, this is the place. The role You will own the research that makes our inference layer more accurate, more reliable and easier to audit, and you will carry that research into production yourself. This is applied work on live financial documents and data, not a lab seat. What you will do:
• Design and run the evaluation harnesses that decide whether a workflow ships: task-level accuracy, failure taxonomies, regression suites, and the evidence a bank's model-risk team asks for.
• Specialise models to individual financial workflows and show they match the frontier model on the same task at a fraction of the cost.
• Improve the extraction and reasoning methods behind multi-document, multi-hop extraction and validation on contracts, term sheets, confirmations and broker feeds.
• Build the monitoring and retraining methods that keep a deployed workflow accurate as documents and counterparties change.
• Benchmark frontier and open-weight models on our workflows and turn the results into clear recommendations on what we run where.




• Work directly with our engineering pods and product leads to move methods from notebook to production, and write them up so clients, auditors and the rest of the team can follow what was done and why. You will see your work in front of Tier-1 clients within months, not years. What we are looking for Must have:
• Hands-on experience fine-tuning, distilling or evaluating LLMs, with results you can point to — not only prompting.
• Strong Python and the modern stack: PyTorch, Hugging Face Transformers, PEFT or similar, vLLM or comparable serving, and the tooling to make experiments reproducible.
• Rigour in experimental design and evaluation: you build the harness before you build the model, and you can explain a result to a sceptical reviewer.
• Solid foundation in statistics, machine learning and software engineering; comfortable with large, messy, unstructured datasets.
• NLP depth: information extraction, document understanding, retrieval and structured-output methods.
• Explicit written and spoken communication; able to work in a small cross-functional pod and own outcomes.
• Master's or PhD in Computer Science, Statistics, Applied Mathematics, Electrical Engineering or a related field, or equivalent practical experience. Strong plus:
• Published work or open-source contributions on evaluation, distillation, structured extraction or neuro-symbolic methods.
• Experience with model governance in a regulated setting: model-risk documentation, validation evidence, lineage.
• Familiarity with financial documents and data — trade confirmations, ISDA or prime brokerage agreements, term sheets, reconciliation breaks, margin calls.
• Experience running models on dedicated infrastructure rather than only on hosted APIs. Location and travel Full-time, based in Pune, with frequent travel to London and New York to work alongside clients' and partners' research teams. Expect regular stretches on site in both cities. You will work alongside senior leaders with 125+ years of combined experience building and selling for financial services, in a profitable company with no pressure to chase a pilot that will never ship.

📌 AI Researcher (Pune)
🏢 Intellimation.ai
📍 Pune