Open to GenAI / ML Engineer roles

Pranav Kakde

Building agentic AI systems, RAG pipelines, and LLM-powered applications that ship to production.

Portrait of Pranav Kakde

About

AI engineering with production intent

Generative AI / Machine Learning Engineer with hands-on experience building agentic AI systems, RAG pipelines, and LLM-powered applications using Python, TensorFlow, Keras, LangChain, LangGraph, Hugging Face Transformers, and Gemini.

Skilled in prompt engineering, vector databases, tool-calling architectures, NLP, and model evaluation — with an applied background in data analysis (Power BI, Excel, SQL). Seeking to contribute to production-grade Generative AI / ML engineering teams building LLM-driven products.

3+

LLM & agentic AI systems built

1.6M+

Tweets processed in NLP pipeline

87.99%

Sentiment model test accuracy

SIH'24

National hackathon finalist

Skills

The stack behind the AI systems

From model training and evaluation to RAG, agents, and deployment.

Generative AI / Agentic AI

  • LangChain
  • LangGraph
  • RAG
  • LLMs (Gemini, Hugging Face)
  • Prompt Engineering
  • Tool Calling
  • Multi-Agent Systems
Proficiency90%

Machine Learning / Deep Learning

  • PyTorch
  • TensorFlow
  • Scikit-learn
  • NLP
  • Transformers
  • Feature Engineering
  • Model Evaluation
  • Classification
Proficiency86%

Vector Search

  • FAISS
  • ChromaDB
  • Embeddings
  • Semantic Search
Proficiency85%

Programming

  • Python
  • SQL
  • Java
Proficiency88%

Data Science

  • Pandas
  • NumPy
  • Matplotlib
  • Seaborn
  • EDA
  • Data Preprocessing
Proficiency88%

Backend & Deployment

  • FastAPI
  • Streamlit
  • Docker
  • REST APIs
  • Git
  • GitHub
  • n8n
Proficiency82%

Databases

  • PostgreSQL
  • MySQL
  • MongoDB
  • SQLite
Proficiency84%

Showing the AI Engineering toolkit — switch above for the other track.

Projects

Agents, pipelines and models, shipped

Each project started with a real problem and ended with something someone could actually use.

AI Customer Support Agent preview
  • LangGraph
  • LangChain
  • Gemini
  • RAG
  • SQLite
  • Streamlit

AI Customer Support Agent

Conversational support agent with stateful workflow orchestration, RAG knowledge base, and LLM tool calling.

  • Built a conversational AI customer support agent using LangGraph, LangChain, and Google Gemini with stateful workflow orchestration, conditional routing, and multi-turn conversation handling
  • Implemented Retrieval-Augmented Generation (RAG) over a support knowledge base covering returns, refunds, shipping, payments, troubleshooting, and warranty queries
  • Integrated LLM tool calling for order status, customer lookup, ticket lookup, and knowledge-base search so the agent performs actions beyond text generation
  • Developed persistent conversation memory with SQLite and thread-based sessions to retain customer context across conversations
  • Built an interactive Streamlit app with conversation history, customer-memory retrieval, tool execution, and escalation for unresolved queries
  • Containerized with Docker and configured environment-based API key management for deployment
GitHub
Twitter Sentiment Analysis preview
  • Python
  • NLP
  • Transformers
  • Scikit-learn
  • Streamlit

Twitter Sentiment Analysis

NLP sentiment-classification pipeline on 1.6M+ tweets reaching 87.99% accuracy across four classes.

  • Built an NLP sentiment-classification pipeline on 1.6M+ tweets with text normalization, stopword removal, and noise filtering
  • Engineered features using TF-IDF and Count Vectorization
  • Benchmarked Logistic Regression and Naive Bayes classifiers for 4-class sentiment prediction
  • Achieved 87.99% accuracy on held-out test data across Positive, Negative, Neutral, and Irrelevant classes
  • Deployed an interactive Streamlit application for real-time prediction and visualization
GitHub
AI Personal Assistant Workflow preview
  • n8n
  • Gemini
  • Docker
  • Google APIs

AI Personal Assistant Workflow

Modular AI agent automating 6 multi-step workflows across email, calendar, notes, and expenses.

  • Designed an AI-powered personal assistant using n8n to automate 6 multi-step workflows, reducing manual handling for email, scheduling, and notes
  • Integrated Gmail, Google Calendar, Google Search, and Notes APIs for end-to-end automation of drafting, scheduling, and retrieval
  • Built a centralized AI Agent that interprets natural language requests and routes them to workflows via webhooks and API calls
  • Connected Gemini/LLMs with memory components to maintain contextual state across tasks
  • Containerized the system with Docker into modular, independently deployable components
GitHub

Achievements

Recognition & certifications

  1. 🏆

    Smart India Hackathon 2024 Finalist — NIT Srinagar

    Geofencing-based attendance system for GAIL · Ministry of Education, Govt of India

  2. ☁️

    AWS Cloud Foundation Badge

    AWS Academy

  3. 🗄️

    SQL Certification

    Oracle Academy

  4. 📊

    Data Analytics Certification

    Simplilearn

Education

Academic background

Master of Computer Applications (MCA)

IICMR, Pune

2025 – Present

Bachelor of Science (B.Sc.) in Computer Science

MIT College, BAMU University — Chhatrapati Sambhajinagar

2022 – 2025

9.09

CGPA / 10

Contact

Let's talk data

Have a dashboard to build, a dataset to untangle, or a role to fill? Drop a message.

This opens your email app with the message pre-filled, ready to send.