Harshal Zarikar

AI Engineer specializing in Agentic AI and Applied Generative AI. Proven ability to architect autonomous workflows using LangGraph and Llama-3, moving beyond basic chatbots to build stateful, self-correcting systems. Hands-on experience with RAG optimization, Docker containerization, and Cloud Deployment (AWS). Passionate about solving engineering challenges like context-window limits and hallucination reduction in production-grade LLM applications.

Generative AI Intern at Finsocial Digital System (2024-11 – Present)

Generative AI Intern working on RAG pipeline development, backend integration, data engineering, and containerization.

• RAG Pipeline Development: Engineered a customer support RAG system that improved query resolution rates by implementing semantic chunking strategies and validating outputs against a DeepEval test set.
• Backend Integration: Optimized real-time chatbot throughput by migrating synchronous endpoints to Asynchronous FastAPI, utilizing Pydantic models for strict data validation and type safety.
• Data Engineering: Automated the collection of unstructured financial data by building robust web scrapers, cleaning and structuring datasets for high-quality Vector Database ingestion.
• Containerization: Standardized the development lifecycle by creating optimized Docker images for backend services, reducing environment setup time by 90% across the engineering team.

B.Tech in Computer Science & Engineering (AI & ML) (2020-06 – 2024-06)