Himanshu Gangwar

AI & Machine Learning Engineer

Baheri, Uttar Pradesh, India

About Me

A dedicated and aspiring AI and Machine Learning engineer with a passion for developing intelligent systems. Experienced in Automatic Speech Recognition (ASR), Natural Language Processing (NLP), and building interactive applications. Proven ability to work with various data models and a strong foundation in core machine learning concepts. Eager to apply my skills to solve real-world problems and contribute to innovative AI projects.

Contact Me Download CV Book a Meet

Technical Skills

Programming Languages

Python

C++

Java

Web Technologies

Streamlit

HTML

LLM Tools & Technologies

LangChain

ChatGroq API

Pinecone

Hugging Face

NLP

Data Visualization

Matplotlib

Seaborn

Plotly

Data Science & Machine Learning

Numpy

Pandas

Scikit-learn

Neo4j

Computer Vision

Deep Learning

MySQL

Tools & Platforms

Git

Docker

Google Colab

Jupyter Notebook

Kaggle

GitHub

VS Code

Experience

AI Intern

Investkode AI | Feb 2026 - Present · Remote

Building a financial GPT assistant to help users analyze markets, portfolios, and company fundamentals.
Designing a concall agent to answer questions from earnings and investor call transcripts.
Designing and iterating on database schemas to support scalable vector search and analytics.

AI Research Fellow · Trainee

Whissle | Jul 2025 - Oct 2025 · California, United States (Remote)

Researched and fine-tuned advanced ASR models tailored for multilingual enterprise use cases.
Ran experimentation pipelines to benchmark accuracy, latency, and robustness across noisy speech datasets.
Partnered with the AI/ML team to productionize promising ASR model variants for customer pilots.

Research Intern · Artificial Intelligence & Machine Learning

Whissle | Mar 2025 - Jun 2025 · Remote

Prototyped new AI models for voice intelligence and evaluated them against established baselines.
Built data augmentation and labeling workflows that improved downstream ASR training efficiency.
Documented research findings and shared weekly insights with senior researchers to guide roadmap decisions.

AI Lead · Part-time

BharatWork | Sep 2025 · Lucknow, Uttar Pradesh (Hybrid)

Architected a custom text-to-speech backend leveraging AWS and MediaPipe signals for BharatWork.
Integrated TTS services with existing product APIs to deliver conversational voice support.
Defined deployment and monitoring strategy to ensure low latency for regional users.

Education

B.Tech in Computer Science and Engineering (AI)

Institute of Engineering and Technology, Lucknow

Sept 2023 - Sept 2027

CGPA: 8.1/10.0

Intermediate (12th Grade)

Shishu Bharti Inter Collage, Rudrapur

Completed May 2023

Grade: 89.6%

High School (10th Grade)

Shishu Bharti Inter Collage, Rudrapur

Completed May 2021

Grade: 85.4

Projects

MCP LeadGen Orchestrator: Agentic Lead Generation & Outreach Automation

Python

Meta Developer API

Google Developer API

FastAPI

Model Context Protocol (MCP)

n8n

SQLite

LLMs

Docker

Architected an agentic B2B sales pipeline using FastAPI, n8n, and Model Context Protocol (MCP) to automate real-time lead capture from Meta Lead Ads API and Google Forms. Developed a local semantic search engine using sentence-transformers and NumPy to dynamically retrieve the most relevant product solutions based on lead industry, role, and stated challenges. Automated highly personalized prospect outreach using LLMs with strict anti-hallucination guardrails, A/B testing logic, and state tracking via SQLite.

GitHub Live Demo

MEDGraphy

Python

Streamlit

LangChain

Neo4j

RAG

A medicine graph RAG-based project that leverages a knowledge graph for medical information retrieval, providing accurate answers to complex medical queries.

GitHub Live Demo

Automatic Number Plate Recognition (ANPR) System

Python

PyTorch

YOLOv8

Gradio

OpenCV

Hugging Face Hub

A complete Automatic Number Plate Recognition (ANPR) system. It uses a fine-tuned YOLOv8 model for plate detection and a custom-built CRNN in PyTorch for character recognition, all wrapped in an interactive Gradio web app.

GitHub Live Demo

Employee NLQ: Production-Grade Natural Language Query System

React

FastAPI (Python)

TypeScript

PostgreSQL

Vector DB (pgvector/FAISS)

LLMs (Groq/OpenAI)

Redis

Docker

Azure

CI/CD

Vercel

Engineered a full-stack NLQ platform with auto-schema discovery, relationship inference, and hybrid query execution across structured (SQL) and unstructured (document) data sources without hard-coded schemas.

GitHub Live Demo

AI-Samvad: Multi-Domain AI Interview Platform

Next.js

FastAPI

TypeScript

WebRTC

PostgreSQL

LLMs

Hugging Face

PlayAI

MediaPipe

Azure

Vercel

Full-stack adaptive AI interview platform for domains like NLP & CV. Real-time audio/video via WebRTC, Whisper STT + multi‑TTS, gaze & browser anti-cheat, automated multi-model (Gemini, Groq, Hugging Face, PlayAI, Whisper) evaluation pipeline with personalized feedback + reporting.

GitHub Live Demo

AI-Powered Research Assistant (MCP Server)

MCP Protocol

Python

FastAPI

JSON-RPC

Arxiv API

GitHub API

PyGitHub

Groq API

LLMs

Custom MCP-compliant research automation server: natural language literature queries, Arxiv search + metadata extraction, Llama/Llama-8B summarization, GitHub publishing workflow, structured Markdown knowledge base, dual HTTP JSON-RPC + stdio MCP interfaces for Claude & compatible clients.

GitHub Live Demo