LLM Scientist

Role: LLM Scientist

(research, design, prototype advanced Generative AI systems)
Focus: Agentic AI design, RAG architecture, LLM fine-tuning, rapid Python prototyping

Key Responsibilities

• Design/prototype Agentic AI / Multi-Agent systems using LangGraph, LangChain, MCP.
• Fine-tune LLMs using LoRA/QLoRA; run controlled experiments (A/B tests, ablation studies).
• Architect and iterate RAG pipelines (chunking, embeddings, hybrid search, reranking).
• Build evaluation frameworks for GenAI (groundedness, hallucination, relevance) and implement observability (e.g., LangFuse).
• Apply statistical rigor: hypothesis testing, confidence intervals, significance testing, sample-size considerations.

Required Skills

• Strong Python coding; build experimental pipelines/tooling from scratch.
• Deep understanding of LLMs, Transformer architecture, and GenAI fundamentals.
• Hands-on experience with Agentic AI, RAG, and LLM fine-tuning.
• Proficiency in LangChain, LangGraph, MCP.
• Experience with vector DBs: FAISS, ChromaDB, Pinecone.
• Strong grasp of ML/stats: distributions, hypothesis testing, bias-variance, regularization, overfitting/underfitting.
• Model evaluation metrics: Precision, Recall, F1, ROC-AUC; classical ML (regression, classification, clustering, ensembles).

Good to Have

• NLP techniques (BERT, embeddings, sentiment analysis).
• Classical ML (XGBoost, anomaly detection).
• Research background/publications in NLP/LLMs.
• Experience in regulated enterprise domains.
• B.Tech/M.Tech in CS/IT or related field.