Vishal Gaikwad

About

AI Engineer focused on building agentic AI systems, LLM applications, RAG pipelines, and AI evaluation systems. Experienced with multi-agent orchestration, tool calling, memory systems, vector search, and backend development using Python, FastAPI, LangChain, and LangGraph.

Education

• BS in Data Science and Applications
  IIT Madras
  Data Science and Applications · Sep 2023 – Dec 2027
• Diploma in Data Science
  IIT Madras
  Data Science · Jan 2025 – Apr 2026

Skills

• FastAPI <3 months
• Python <3 months
• Power BI <3 months
• Prompt Engineering <3 months
• Feature Engineering <3 months
• Model Evaluation <3 months
• scikit-learn <3 months
• PyTorch <3 months
• AWS <3 months
• CNN (Convolutional Neural Network) <3 months
• LangChain <3 months
• Databricks <3 months
• API Integration Understanding <3 months
• RAG <3 months
• Vector Search <3 months
• Multi-Agent Systems <3 months
• LangGraph <3 months
• MCP <3 months
• LangSmith <3 months
• Embeddings <3 months
• SQLAlchemy <3 months
• Transfer Learning <3 months
• Ragas <3 months
• Chromadb <3 months
• CrewAI <3 months

Tools / apps / platforms

• Docker 3 to 6 months
• Git 6 to 12 months
• PostgreSQL 6 to 12 months
• Redis 3 to 6 months
• SQLite 6 to 12 months

Languages

• English Fluent Speak · Read

Projects

• Compass — Proactive Growth Opportunity Engine
  Anthropic SDK, FastAPI, PostgreSQL, DuckDB, APScheduler

  Built an agentic analytics system with scheduled analysis and chat-based interaction connected to a shared analytics core. Developed parallel detectors for anomalies, high-value segments, and historical patterns using validated read-only SQL queries. Added SQL injection protection, PII masking, human-in-the-loop controls, and CI-based evaluation for precision, recall, and answer faithfulness.

• Image Memory — Multimodal AI Memory System
  FastAPI, Streamlit, OpenRouter, SQLite, sqlite-vec, OpenCV, Python

  Designed a modular FastAPI system for image ingestion, semantic indexing, vector retrieval, and smart collections. Built a multimodal pipeline that generates image captions, metadata, semantic embeddings, and facial embeddings. Implemented semantic search, reverse image search, duplicate detection, and related-image recommendations with an interactive Streamlit dashboard.