We are seeking a skilled developer to design, optimize, and deploy offline/on-premise generative AI models and LLMs. This role focuses on building high-throughput, low-latency inference systems and robust MLOps pipelines for private environments.

Responsibilities:
Deploy and optimize LLMs using frameworks like TGI, vLLM, DeepSpeed
Accelerate models with TensorRT-LLM, quantization, GPU optimizations
Build offline RAG/agent systems with LangChain, Google ADK, SgLang
Fine-tune models (LoRA/QLoRA) for edge and constrained environments
Develop CI/CD pipelines with Docker & Kubernetes/OpenShift
Create production-ready APIs with FastAPI, SQLAlchemy/Alembic
Research recent inference algorithms & quantization techniques

Qualifications:
Bachelor’s/Master’s in CS, Engineering, or a related field
4–5 years ML experience (with minimum 2 years of experience in LLM optimization/deployment)
Strong Python & CUDA/GPU expertise
Proven record of deploying offline, high-performance LLMs

Preferred:
Experience in air-gapped/non-cloud environments
Knowledge of Spark/Kafka for RAG pipelines
Contributions to open-source LLM frameworks

📌 GenAI and Open-source LLM Developer (Mumbai)
🏢 L&T Precision Engineering and Systems
📍 Mumbai