Om Nagvekar

About Me

Hello, my name is Om Nagvekar

I’m exploring AI Research and Agentic AI Engineering roles

Om Nagvekar

Having previously worked as a Data Science Intern at AlgoAnalytics, I gained hands-on experience in quantitative financial statistics, modeling, and NLP-driven analytical systems. My work involved building real-world evaluation frameworks such as walk-forward and experimental model validation, along with generating data-backed insights for applied problem-solving.

I have also contributed to successful copyright filings of RAG-based research tools and Hysteresis change-point detection software at Shivaji University, developed synthetic financial data pipelines at ArthaVedh (ArthaVedh Consulting), and built contract intelligence systems using fine-tuned LLMs during Intel’s Unnati program. Earlier, I co-authored a research publication on spam image detection in Android galleries, published by Springer in 2024.

My current focus lies in Agentic AI, autonomous agents, knowledge-graph assistants, computer vision, and scalable ML systems. I enjoy turning research ideas into working, production-ready systems whether as APIs, reasoning agents, or graph-powered assistants, including projects like SIH 2024’s Garbage Prediction API and the Chef-Agent knowledge graph assistant.

I’m passionate about working at the intersection of AI research and real-world impact, especially in agent-based reasoning systems and applied ML. I’m always exploring new frameworks, automation-driven AI, and scalable ML-Ops practices to build systems that are both intelligent and practical.

Let’s connect feel free to reach out via email, LinkedIn, or the contact form. I’d be happy to chat, collaborate, or explore AI research opportunities!

Download Resume

Education

KIT’s College of Engineering (Autonomous), Kolhapur

Dec 2021 – Jun 2025

B.Tech in Computer Science (Artificial Intelligence and Machine Learning)
GPA: 8.41/10

Sou. S.M. Lohia Jr. College, Kolhapur

2019-2021

12th (HSC, Maharashtra State Board)
Percentage: 88.33%

New High School, Kolhapur

2019

10th (SSC, Maharashtra State Board)
Percentage: 82.80%

Experience

Data Scientist Intern, AlgoAnalytics Private Limited

Aug 2025 – Dec 2025
  • Working on Quantative Financial Statistics, Data Analysis and Data Modeling.
  • Calculated the alpha-beta value for different stocks to evaluate their performance.
  • Developed a Value–Momentum-based trading strategy, achieving a 32% Annualized Return with an Annualized Volatility of 25% over a 10-year backtest on NSE stocks.
  • Implemented convex portfolio optimization with covariance-driven risk control.

Research Intern, Shivaji University (Nanoscience Department)

Aug 2024 – Jun 2025
  • Worked under the guidance of Dr. T. D. Dongale and Dr. S. S. Sutar.
  • Applying machine learning techniques (time series analysis and change point detection) to material science data.
  • Developing a RAG-based system for research paper analysis to extract key scientific insights, potentially increasing data extraction efficiency by 50%.
  • Secured 2 copyrights for innovative contributions in applied machine learning and scientific data analysis.

Machine Learning Intern, ArthaVedh Consulting

Jun 2024 – Dec 2024
  • Contributed to the development of machine learning models and synthetic data generation, improving training workflows by streamlining data pipelines by over 30%.
  • Built synthetic data generators with Faker & pandas.

Intern – Business Contract Validation, Intel Unnati Industrial Training Program 2024

May 2024 – Aug 2024
  • Automated validation of hundreds of business contracts by classifying and comparing them against standard documents, reducing manual review time by an estimated 40%.
  • Implemented a RAG pipeline with a fine-tuned Phi-3 mini LLM to accurately identify deviations.

Skills

Programming Languages

Python, Java

AI & ML Frameworks

PyTorch, TensorFlow, Keras, LangGraph, LangChain, smolagents, FastMCP

NLP & RAG

Retrieval-Augmented Generation (RAG), LangChain, LlamaIndex, LLM fine-tuning

API Development & Tools

FastAPI, Docker, Model Context Protocol (MCP)

Research & Data Processing

Data preprocessing, exploratory analysis, model evaluation, pandas, numpy

Version Control

Git, GitHub

DevOps

Docker, Kubernetes, AWS

Languages

English

Japanese

Other

Android Development, SQLite

Publications

Managing Spam Images on Android: An Approach Utilizing Machine Learning and NLP

Om Ulhas Nagvekar, Sumeet Kurbetti, Parth Sarnobat, Uma Gurav, and Tanvi Patil.
Proceedings of Fifth International Conference on Computing, Communications, and Cyber-Security: IC4S’05, Volume 1, Springer, Lecture Notes in Networks and Systems.
DOI: https://doi.org/10.1007/978-981-97-2550-2_59

Certifications and Awards

Projects

Information Retrieval RAG for Research Papers

  • Built a RAG-based system to extract structured insights from research papers using HuggingFace, Ollama, and Gemini API.
  • Designed dynamic schema extraction with runtime Pydantic models and integrated citation-linked outputs.
  • Developed a Streamlit UI with PDF uploads, schema CRUD, and MongoDB-backed multi-user chat history
  • reducing manual processing by 40%.

LINK

Business Contract Validation

  • Built an automated tool for classifying and validating contracts using a TensorFlow model, RAG pipeline, and HuggingFace embeddings.
  • Integrated Gemini AOI (Gemini 2.0-Flash) for clause-level evaluation and Streamlit UI for PDF visualization.
  • Achieved 92.5% semantic similarity, 88.7% response relevancy, and 85.3% factual correctness across test cases.
  • Tools: Python, TensorFlow, LlamaIndex, Gemini AOI, Streamlit, ragas.

LINK

Chef-Agent Knowledge-Graph Cooking Assistant

Designed and implemented a streaming AI “Chef” agent using FastAPI/FastMCP, LangGraph workflows, LangChain and a Neo4j‑backed recipe knowledge graph to support real‑time recipe querying, web scraping, dynamic graph updates, and personalized memory.

  • Developed a streaming AI “Chef” agent using FastMCP + FastAPI, orchestrated via LangGraph workflows and backed by a Neo4j recipe knowledge graph.
  • Implemented end-to-end tools for web search (Tavily/DuckDuckGo), web scraping (FireCrawl + BeautifulSoup), sandboxed Python execution, natural-language → Cypher graph queries, and URL-to-graph recipe ingestion.
  • Enabled session personalization with in-memory/Redis store and auto-summarization of long conversations for enhanced user experience.

LINK

GAN Implementations

Implemented GAN variants using PyTorch and research papers, optimizing training stability and image quality.

LINK

AI_Predictions_API

A FastAPI project providing endpoints for analyzing images and videos to predict garbage intensity, types, and other characteristics using advanced deep learning models.

  • Designed a FastAPI-based API to analyze images and videos in real time for garbage prediction.
  • Integrated advanced LLMs (Phi-3 Mini and Florence-2) to classify and quantify waste, improving monitoring efficiency by 20%.

LINK

SpamSnap

An Android application to manage redundant photos in your gallery.

LINK

Blog

Loading posts…

Notes

Check out my collection of PDF and document notes covering a wide range of topics including Cloud, Android, Docker, Git, Kubernetes, Machine Learning (ML), Artificial Intelligence (AI), Data Structures & Algorithms (DSA), and more.

View Notes

Contact