AI Engineer

Role: AI Engineer\nLocation : Gurugram, Haryana, India\nWorkplace type: Hybrid i.e. 6 days a week (2-3 days per week in-office, rest remote)\n\nAbout the Role\nWe are looking for an AI Engineer who can build production-grade AI systems—not just connect applications to third-party APIs.\n\nYou will work across the AI development lifecycle: selecting and fine-tuning models, designing multi-agent architectures, building evaluation pipelines, deploying inference systems, and improving model performance using real-world feedback. This role is ideal for an engineer who enjoys experimentation but also knows how to turn prototypes into reliable, scalable products.\n\nExperience working in LegalTech, compliance, finance, healthcare, or another domain involving sensitive and complex data is highly preferred.\n\nResponsibilities\nDesign, build, and deploy LLM-powered applications and intelligent workflows.\nFine-tune and adapt open-source models using techniques such as LoRA, QLoRA, SFT, and preference optimization.\nBuild multi-agent systems with effective planning, tool usage, memory, routing, and human-in-the-loop controls.\nDevelop RAG pipelines, including document processing, embeddings, retrieval, reranking, and citation generation.\nCreate evaluation frameworks for accuracy, hallucination, latency, cost, safety, and task completion.\nOptimize inference performance and deploy models in cloud or self-hosted environments.\nWork with product, engineering, and domain experts to translate complex legal or business workflows into AI systems.\nMonitor production performance and continuously improve models using user feedback and evaluation data.\n\nRequirements\n2+ years of experience in AI, machine learning, NLP, or applied LLM engineering.\nStrong Python and software engineering fundamentals.\nHands-on experience with PyTorch, Hugging Face Transformers, or similar frameworks.\nExperience fine-tuning, evaluating, and deploying language models.\nUnderstanding of transformer architectures, embeddings, tokenization, context management, and inference optimization.\nExperience building agentic or multi-agent systems using frameworks such as LangGraph, AutoGen, CrewAI, or custom orchestration.\nFamiliarity with vector databases, structured outputs, tool calling, and retrieval systems.\nAbility to independently take an AI feature from experimentation to production.\n\nPreferred Qualifications\nExperience at a LegalTech startup or working with legal documents, contracts, case law, compliance, or document review.\nExperience with model serving tools such as vLLM, TGI, Triton, or Ollama.\nFamiliarity with LLM observability, guardrails, prompt-injection risks, and data privacy.\nExperience working in a fast-moving startup environment.\n\nWhat We Value\nWe value engineers who understand models deeply, question default approaches, run disciplined experiments, and build AI systems that are accurate, explainable, secure, and useful in production.