Data Scientist

JD:
Reporting Structure
Project Manager
Education
Bachelor's Degree in Engineering
Experience
3 years in applied ML/AI/Gen AI lead roles with demonstrable production deployments; Mandatory experience with on-prem LLMs and model hosting, including fine-tuning and inference optimization. Proven expertise with Vector Databases and embedding pipelines for retrieval augmented generation and semantic search. Strong statistical foundations and experience designing experiments and evaluation frameworks. Proficiency in Python, PyTorch or TensorFlow, and model serving frameworks. Familiarity with containerization and orchestration for on-prem environments. Proven experience and knowledge on LLM fine-tuning, embeddings, RAG, transformers. Hands-on experience with advanced RAG architectures, including hybrid search, semantic search, reranking, metadata filtering and retrieval optimization. Experience on Vector DBs: Elastic, Milvus, etc. and on indexing and similarity tuning. Experience evaluating retrieval performance using Precision@K, Recall@K, Mean Reciprocal Rank (MRR), NDCG and Hit Rate@K. Experience with Agentic AI, tool/function calling, multi-step workflows and orchestration frameworks such as LangChain and LangGraph. MLOps: model registries, CI/CD, automated testing, monitoring and drift detection. Experience with model explainability tools, quantization, and inference acceleration. Familiarity with OpenShift or enterprise Kubernetes for on-prem deployments.
Industry
Preferably from a Technology, ITES, Products, Professional SI organization with a focus on BFSI.
Responsibilities

• At least one production implementation of an AI, ML or Gen AI application with actual implementation experience and knowledge of various production grade application being used by End users • Define and track model metrics: accuracy, precision/recall, AUC, calibration, latency, throughput, business KPIs, and translate them into SLOs/SLA targets. • Design and operate on-prem LLMs: fine-tuning, prompt engineering, RAG architectures, embedding generation, and secure inference. • Integrate and manage Vector Databases: design embedding pipelines, select indexing strategies, tune similarity metrics, implement secure access controls, and monitor vector DB performance and cost. • Design and implement end-to-end RAG evaluation pipelines to measure retrieval relevance, answer correctness, faithfulness and hallucination rates. • Implement hybrid retrieval strategies combining keyword search, vector search and reranking to improve information retrieval accuracy. • Implement robust validation and governance: bias and fairness checks, explainability, model risk assessments, drift detection, and automated retraining triggers. • Productionize models with CI/CD, model registries and automated validation gates. • Collaborate with stakeholders across product, platform, security, and compliance to align models with business needs and regulatory constraints.

Certifications
Data Science, Data Analytics, Big Data Certifications are preferred