Harsh is a Machine Learning Engineer with 10+ years of experience specializing in AI and GenAI solutions. He has a proven track record of leading end-to-end ML project lifecycles and deploying models on cloud platforms.
Led AI engineering for complex query answering solutions using Langgraph.
Engineered and deployed ML models utilizing Docker and Kubernetes.
Applied advanced optimization techniques for logistics problems across 30,000+ locations.
Successfully finetuned RAG approaches with LLMs for enhanced document search.
Developed and deployed an AI-based agent solution for complex query answering.
Designed an efficient BI Dashboard for cloud-based kitchen data analysis.
Overview: Developed an AI-based agent solution to answer multiple-composite queries using Langgraph, generating concise summaries and invoking other functions. Responsibilities: Led AI engineering, design, and finetuning of the OpenAI platform.
Key outcomes:
Successfully developed an agent-based approach for complex query answering.
Overview: Developed an AI solution, CharacAI, to help end-users get probable diseases and solutions based on diagnosis, leveraging LLMs to find context-related articles and texts for medical practitioners. Responsibilities: Focused on AI engineering, design, and development of a RAG pipeline using Azure RAG.
Key outcomes:
Developed a POC for a medical diagnosis AI system.
Overview: Developed a platform to provide used engineering books at affordable rates. Responsibilities: Managed end-to-end design, development, and deployment. Utilized AWS EC2, Spark for analysis, and Docker for deployment.
Key outcomes:
Developed and deployed an e-commerce platform for used books.
Cluster similar company names — 3 million similar company name clustering with limited resources using Whoosh Indexing + Kubernetes + Azure DevOps + Text Vectorization as Team Lead + Solution Architect.
Key outcomes:
Successfully clustered 3 million company names with optimized resource usage.
Led the development and deployment of the solution.
Yard Management System — optimal container placement in 30,000+ locations minimizing shuffling/retrieval costs with Genetic Algorithm + ECS as Solution Architect + Team Lead.
Key outcomes:
Successfully optimized container placement in over 30,000 locations.
Reduced shuffling and retrieval costs using advanced optimization algorithms.
Harsh
ML Ops Engineer