Hi, I'm

|
400+
DSA Problems
8.45
CGPA
6
Projects

About Me

I am Nishit Kumar, a final-year B.Tech student at the National Institute of Technology Karnataka (NITK), Surathkal, with a strong focus on Data Science, Machine Learning, and Generative AI.

My work centers on Generative AI and Agentic AI systems, where I design and build multi-agent pipelines that translate complex reasoning into reliable, production-ready applications.

My technical foundation spans Python, SQL, and core ML/DL frameworks (TensorFlow, Keras, Scikit-learn), extending into the GenAI stack — Hugging Face Transformers, FAISS, ChromaDB, SentenceTransformers, and the Groq API — along with a solid grounding in statistical modeling, hypothesis testing, and feature engineering.

Beyond technical work, I've contributed to campus initiatives as a Marketing Executive for Incident, NITK's annual cultural fest, leading outreach to 10+ sponsors, and as a Web Developer for the Hindi Evam Sanskrit Club, maintaining a platform serving 500+ members. I ranked in the top 3% of JEE Advanced/Mains among over a million candidates, and have solved 400+ problems across LeetCode and GeeksforGeeks — a habit that keeps my problem-solving and algorithmic thinking sharp.

Nishit Kumar

Featured Projects

Showcasing my best projects and real-world work

Smart Shopping Agent

Agentic AI shopping assistant using LangGraph, Groq and SerpAPI

  • 7-node multi-agent pipeline with intent classification and multi-turn memory
  • Live product scoring, fake discount detection and hallucination-proof output
  • Deployed on Streamlit Cloud
Python · LangGraph · Groq · SerpAPI · Streamlit · Pydantic

HireMind

Intelligent candidate screening and interview orchestration using multi-agent workflows

  • Semantic resume analysis and candidate matching with Chroma vector store
  • Automated job description parsing and contextual interview question generation
  • Complete pipeline with PDF report generation and interactive UI dashboard
Python • LangGraph • LLaMA 3 • SerpAPI • Groq API • Streamlit

Consumer Complaint Classifier

Deep learning NLP system for classifying financial complaints with explainable AI

  • BiLSTM, LSTM & CNN architectures with GloVe embeddings
  • Occlusion-based XAI token highlighting for transparency
  • Automated PDF compliance reporting
Python • TensorFlow • Keras • Streamlit • GloVe • ReportLab

Fashion Recommendation System

Recommends visually similar fashion items using deep learning

  • ResNet50-based image embeddings
  • FAISS similarity search on 44K+ items
  • Fast real-time recommendations
Python • PyTorch • ResNet50 • FAISS • TensorFlow

Candidate Calling Bot

Automates candidate calling and response collection using voice-based AI

  • Automated calling with predefined questions
  • Voice input collection using Twilio API
  • Real-time data logging with Google Sheet
Python • Flask • Twilio API • Google Sheets API • JavaScript

Loan Approval Prediction System

Predicts loan approval using machine learning on applicant data

  • Multiple models (RF, SVM, KNN) for comparison
  • Optimized performance using GridSearchCV
  • End-to-end pipeline with web deployment
Python • Flask • Scikit-learn • Pandas • NumPy • Matplotlib

Skills & Technologies

Technical expertise across the ML stack

Programming & Core

Python C++ SQL

Frameworks & Tools

Git GitHub Streamlit FastAPI VS Code Jupyter Notebook

AI / ML Libraries

NumPy Pandas Scikit-learn TensorFlow Keras Matplotlib Seaborn

Concepts

Machine Learning Deep Learning NLP Generative AI LLMs Statistical Modeling Hypothesis Testing A/B Testing Feature Engineering EDA Text Embeddings Semantic Search RAG

GenAI & Agentic

LangChain LangGraph FAISS Hugging Face Transformers ChromaDB SentenceTransformers Pydantic Groq API

Experience & Activities

June 2024 – Mar 2025

Marketing Executive

Incident - NITK Annual Cultural Fest

Leading marketing initiatives for one of South India's largest college cultural festivals. Developed digital marketing strategies and managed social media campaigns reaching 50,000+ students.

August 2024 – Present

Web Developer

Hindi Evam Sanskrit Club (HESC), NITK

Designed and maintained the club website, implementing responsive designs and improving user engagement. Managed technical aspects of online events and competitions.

Get In Touch

Let's Work Together

I'm currently seeking full-time Data Science and Genai Engineering roles for 2027 graduates. Whether you have an opportunity, want to collaborate on a project, or just want to connect, I'd love to hear from you!