About this role
Atlys is hiring a Research Scientist for its AI team at its Delhi headquarters. The role focuses on inventing methods that make AI systems for travel and visa applications more accurate, reliable and autonomous. Research should translate into production improvements in approval rates, turnaround times and cost per application.
Responsibilities
- Set the research agenda for LLM-powered agents that plan, use tools, act and recover from errors during long, multi-step tasks, including work across unfamiliar government portals.
- Develop multimodal methods to extract, verify and reason over noisy, unstructured documents and other real-world inputs across formats and languages.
- Research post-training techniques, including fine-tuning, preference optimisation and reinforcement learning, as well as calibration, uncertainty estimation and self-verification to determine when systems should act or defer to a human.
- Build evaluations, benchmarks and synthetic-data methods that reveal model quality, robustness and long-tail failures. Run careful experiments, collaborate with research engineers to move prototypes into production, and share findings internally or through publications where appropriate.
Ideal candidate:
- A PhD or MS in Computer Science, Machine Learning or a related field, or equivalent research experience, plus 3+ years of applied or academic research in machine learning, NLP or computer vision.
- Original research demonstrated through leading-venue publications, influential open-source work or measurable production impact. Hands-on LLM experience beyond prompting, including fine-tuning, RLHF/DPO or other post-training methods and evaluation design.
- Strong Python and PyTorch skills; ability to conduct data curation, training runs and rigorous analysis independently.
The position is full-time and on-site at Delhi HQ; applicants are asked whether they are based in Delhi NCR or open to relocating. Atlys offers competitive compensation and equity, family health insurance, preventive health checks, paid time off and company holidays, office meals, learning and development credits, relocation support, and research equipment, software and compute. No salary amount is stated.