Bala S
AI/ML Engineer
Hire this person
Send a job offer directly to this candidate
About
Machine Learning / AI Engineer with 12+ years of experience in software development, machine learning, NLP, data engineering, and production AI application development across healthcare, banking, and retail domains. Strong hands-on experience developing end-to-end ML solutions using Python, Scikit-learn, Pandas, NumPy, PySpark, SQL, TensorFlow, and PyTorch. Experienced in Generative AI and LLM application development using AWS Bedrock, LangChain, LangGraph, RAG, embeddings, vector databases, prompt engineering, semantic search, and structured outputs.
Hands-on experience building RAG pipelines covering document ingestion, preprocessing, chunking, embedding generation, vector retrieval, prompt orchestration, LLM processing, and response generation. Strong background in supervised and unsupervised machine learning, including classification, regression, clustering, ensemble methods, feature engineering, model evaluation, and hyperparameter tuning. Experienced in NLP and transformer-based applications using Hugging Face Transformers, BERT, Sentence Transformers, text classification, embeddings, semantic search, and natural-language processing.
Proficient in data preparation and analytics, including EDA, data cleansing, preprocessing, feature engineering, statistical analysis, data-quality validation, and development of ML-ready datasets. Experienced with AWS and Azure ML platforms, including AWS SageMaker, S3, EC2, IAM, Azure Machine Learning, Azure Blob Storage, and cloud-based ML deployment environments.
Developed
FastAPI and Flask REST APIs to integrate machine learning, NLP, and GenAI capabilities with enterprise applications and microservices. Hands-on experience with MLOps and ML lifecycle management using MLflow, experiment tracking, model versioning, CI/CD, Docker, Kubernetes, automated testing, and reproducible training and inference workflows. Experienced in designing autonomous AI agent workflows using Python, LangGraph, LangChain, MCP, and tool/function calling to support multi-step reasoning, enterprise system integration, and automated workflow execution.
Hands-on experience implementing AI safety, prompt optimization, prompt versioning, guardrails, structured outputs, and human-in-the-loop validation to improve LLM reliability, consistency, and controlled execution of AI-driven workflows. Strong understanding of machine learning model optimization, including supervised and unsupervised learning, clustering, classification, feature engineering, hyperparameter tuning, and model evaluation using Scikit-learn and PyTorch. Experienced with vector search and retrieval technologies including Pinecone, FAISS, Qdrant, and semantic-search architectures for AI-powered information retrieval.
Strong software engineering foundation in Python, SQL, OOP, REST APIs, microservices, Git, Jenkins, GitHub Actions, Docker, and Kubernetes, supporting the development of production-ready AI/ML applications. Experienced in implementing CI/CD pipelines and DevOp
Experience
GenAI / ML Engineer at CVS Health (2025 – Present)
- Developed GenAI-powered healthcare insurance application components using Python, LangChain, AWS Bedrock, RAG, embeddings, and vector databases to support claims processing and prior-authorization workflows, enabling contextual information retrieval and reducing manual information-review effort.
- Contributed to conversational AI and agentic workflows for healthcare prior-authorization processes, supporting multi-turn interactions, information gathering, contextual question answering, document retrieval, and guided processing of authorization-related requests.
- Implemented multi-step conversational agent workflows using Python and LangGraph to manage interaction states, maintain conversational context, identify missing information, request clarification, validate required inputs, and guide agent execution through predefined healthcare business workflows.
- Incorporated Human-in-the-Loop (HITL) validation and approval checkpoints into agentic workflows to support controlled execution of authorization-related activities, review AI-generated recommendations, and maintain appropriate human oversight.
- Implemented AI guardrails, workflow controls, and response-validation mechanisms using Python and LangGraph to validate structured LLM outputs, restrict unsupported responses, control tool invocation, and enforce predefined
- application and business rules.
- Contributed to conversational response-handling mechanisms covering incomplete inputs, retrieval failures, un-supported requests, and agent execution errors using Python, LangGraph, and exception handling to improve conversational reliability and workflow consistency.
- Built document ingestion pipelines using Python, REST APIs, and Amazon S3 to process healthcare insurance policies, claims information, clinical documentation, and supporting documents for downstream document processing and retrieval workflows.
- Utilized Databricks, PySpark, and Python to process, transform, cleanse, validate, and prepare large-scale healthcare datasets for downstream analytics, claims processing, and GenAI application workflows.
- Used Pandas and NumPy for data exploration, cleansing, normalization, and metadata preparation to suppo