Back to Careers
Engineering
Sr. AI/ML Engineer
India · RemoteFull-time / Contract4+ years
Take models from idea to production: fine-tuning, evaluations and inference optimisation for legal work.
About Lejit
Lejit is an AI-native law firm with same-day turnaround, enabled by AI and lawyers. Lejit AI runs a law firm's entire client journey. Our all-in-one platform provides outcome-based, full-stack legal services across four areas:
- Tech for law firms
- Legal work for businesses
- AI-native law firm
- Tech for in-house teams
The Role
At Lejit, we need a Sr. AI/ML Engineer to take models from idea to production: fine-tuning, evaluations and inference optimisation for legal work.
What You Will Do
- Fine-tune open models such as Llama, Mistral and Qwen with LoRA/QLoRA/PEFT on legal tasks, and know when deeper tuning is needed.
- Design evaluation harnesses: golden sets, LLM-as-judge with agreement checks against lawyer reviewers, regression gates in CI, and task metrics.
- Optimise inference with vLLM, quantisation, prefix caching and speculative decoding.
- Build RAG and agent pipelines on messy real-world legal data (scanned PDFs, contracts, court orders), and make fine-tune vs prompt vs RAG decisions with numbers.
- Build for confidentiality, so models can run in privacy-sensitive and client-controlled environments.
What We Need
- 4+ years of ML engineering.
- PyTorch and strong Python.
- Production deployment experience.
- Ownership of inference infrastructure, including on-call.