Pranav Gopal Karande

About

Electronics & Telecommunication Engineering student specializing in Generative AI, NLP, and Deep Learning. Experienced with RAG, autonomous AI agents, and AI automation using LangChain, LangGraph, Hugging Face, AWS, and Streamlit.

Experience

• AI Intern
  Edunet Foundation (IBM SkillsBuild)
  Jul 2025 – Aug 2025

  Developed a conversational AI travel planner using IBM Watson Assistant and IBM Cloud. Integrated Maps, weather, flight, and hotel APIs for personalized travel recommendations. Implemented NLP intent recognition and budget estimation for itinerary planning.

• Cloud DevOps Intern
  Cloudnautic
  Jun 2024 – Aug 2024

  Deployed and managed AWS EC2 instances with security configurations and monitoring. Managed Amazon S3 storage and access control. Containerized applications using Docker for consistent deployment.

Education

• B.Tech
  MIT Academy of Engineering, Pune
  Electronics & Telecommunication Engineering · 2023 – 2025

Skills

• SQL <3 months
• Computer Vision <3 months
• Deep Learning <3 months
• FastAPI <3 months
• Python <3 months
• Machine Learning <3 months
• Prompt Engineering <3 months
• PyTorch <3 months
• LLM evaluation <3 months
• LangChain <3 months
• Workflow Automation <3 months
• Natural Language Processing <3 months
• LLM Applications <3 months
• RAG <3 months
• Multi-Agent Systems <3 months
• LangGraph <3 months
• AI Agents <3 months
• MCP <3 months
• Embeddings <3 months
• Ragas <3 months
• Faiss <3 months
• APIs and integrations <3 months
• LoRA <3 months
• Human-In-The-Loop (HITL) <3 months
• Generative AI (ChatGPT, Copilot, Gemini) <3 months

Tools / apps / platforms

• Amazon Elastic Compute Cloud EC2 3 to 6 months
• Amazon Web Services AWS 3 to 6 months
• Docker 3 to 6 months
• Git <3 months
• GitHub <3 months
• Hugging Face <3 months
• Hugging Face Transformers <3 months
• IBM Cloud <3 months
• n8n <3 months
• Pytest <3 months
• S3 3 to 6 months

Projects

• Multi-Lead ECG Disease Classification using InceptionV3-CapsNet
  PyTorch, CapsNet, Dynamic Routing

  Developed a 13-lead ECG image classification pipeline with lead separation and preprocessing. Designed InceptionV3-CapsNet with dynamic routing, lead attention, and multi-lead feature fusion. Achieved 88.54% accuracy and 94.49% ROC-AUC on 253 held-out ECG reports.

• RepoPilot: Autonomous AI Software Engineer ↗
  LangGraph, MCP, Python, FastAPI, Pytest, Docker

  Built a 14-node autonomous agent using LangGraph and MCP to triage GitHub issues and generate verified code patches. Developed an isolated sandbox and self-healing debug loop using Pytest tracebacks, achieving 100% pass rate across 76 tests. Implemented human-in-the-loop gates, pre-commit validation, background daemons, and Discord/Slack alerts.

• Hybrid RAG Knowledge Assistant ↗
  Python, LangChain, FAISS, Streamlit, RAGAS

  Built a citation-aware RAG system using LangChain, FAISS, and Hugging Face embeddings for document-grounded QA. Engineered hybrid retrieval combining semantic and keyword search to improve context relevance. Evaluated with RAGAS and deployed on Streamlit Cloud.

Courses & certifications

• Deep Learning for Natural Language Processing · NPTEL