Sr. MLOps Engineer

Location: India, Remote

Work Experience: 5+ Years

Requirements:

• Strong Python programming.
• Solid understanding of LLMs, transformers, embeddings, and tokenization.
• Experience with RAG systems and retrieval optimization.
• Knowledge of chunking strategies and embedding techniques.
• Hands-on with vector databases (FAISS, Pinecone, Weaviate, Chroma).
• Experience with LangChain / LlamaIndex / LangGraph / similar frameworks.
• API development using FastAPI / Flask.
• Strong data engineering skills (ETL, preprocessing, unstructured data).
• Experience evaluating models (precision@k, recall@k, LLM metrics).
• Familiarity with RAG evaluation tools (RAGAS, TruLens, DeepEval).
• Experience with cloud platforms (AWS / GCP / Azure).
• Knowledge of Docker, Kubernetes, CI/CD pipelines.
• Understanding of data versioning, lineage, and governance.
• Experience with monitoring, logging, and observability.
• Exposure to multi-agent systems and orchestration.
• Knowledge of fine-tuning / LoRA.
• Awareness of Responsible AI and compliance practices.

Qualifications: Bachelor’s degree in computer science, Engineering, Information Systems, or related technical field (or equivalent practical experience).

Job Description:

• Build and deploy AI applications: copilots, chatbots, code assistants, document agents, and decision automation systems. Design and optimize RAG pipelines (chunking, embeddings, retrieval tuning).
• Implement multi-agent systems and orchestration workflows.
• Develop pipelines for prompt data, grounding, and retrieval datasets.
• Build and maintain document loaders and preprocessing pipelines.
• Maintain audit trails for model versions, prompts, datasets, embeddings, and outputs.
• Support CI/CD pipelines for ML code, models, and infrastructure (GitHub‑based).
• Ensure ML workflows are:
  • Reproducible.
  • Traceable.
  • Auditable (aligned with automotive engineering expectations).
• Track data lineage across ingestion → transformation → inference. Ensure reproducibility of ML/LLM pipelines.
• Implement explainability (SHAP, LIME, prompt tracing).
• Enforce access controls and data security policies.
• Align with regulatory standards (GDPR, SOC2, Responsible AI.
• Ensure high data quality (cleaning, parsing, enrichment).
• Implement data versioning, lineage, and governance for AI systems.
• Manage embedding lifecycle (indexing, re-embedding, updates).
• Ensure RAG data freshness via incremental ingestion and re-indexing.
• Evaluate embedding models and retrieval performance.
• Measure LLM output quality (relevance, hallucination, faithfulness).
• Implement robust observability for AI/ML/LLM systems (latency, drift, reliability, performance).
• Track hallucination rates, response quality, and latency.
• Monitor token usage, cost, and prompt effectiveness.
• Build monitoring, observability, and feedback loops.
• Design and optimize prompt templates and system instructions.
• Optimize systems for accuracy, latency, and cost.