A

Hi, I'm Atharv!

I make AI things happen

About Me

I am a Computer Science and Applied Mathematics graduate from IIIT Delhi, with hands-on experience in machine learning, natural language processing, data modeling, and large language model evaluation. My work spans both research and industry. Currently, I am contributing to PixorAI as founding AI engineer, where I evaluate diffusion models, and ensure alignment through rigorous evaluation frameworks.

I enjoy bridging theory with practical impact, whether it’s optimizing training pipelines from hours to minutes, or designing systems that bring clarity to complex models and agentic workflows. Outside of research and development, I love finding playful edges to otherwise serious problems, which is why I like to say: I can debug neural networks faster than I can decide what to eat for dinner.

Work Experience

My Work

Wide and Deep with Multihead Attention (DL Competition)
Wide and Deep with Multihead Attention (DL Competition)

This competition was a part of the coursework for Deep Learning (CSE641) at IIIT Delhi, taught by Dr. Vinayak Abrol in the Winter 2025 semester.

  • Rank: 4th out of 74 participants
  • Final Error on Test Set: 3.9981
Deep Learning
Regression
Attention Mechanism
Model Optimization
Tabular DL
PyTorch
LLM-Powered Policy Impact Assessment Chatbot

Multi-threaded LLM chatbot with chat resume and session persistence. Implemented tool-calling agents for internet search, and RAG pipelines.

  • Implemented agent tracing, to moniter latency and inference costs at each node.
LangGraph
LangChain
LangSmith
OpenAI API
Streamlit
MultiUtility Chatbot

Multi-threaded LLM chatbot with chat resume and session persistence. Implemented tool-calling agents for internet search, and RAG pipelines.

  • Implemented agent tracing, to moniter latency and inference costs at each node.
LangGraph
LangChain
LangSmith
OpenAI API
Streamlit
Conditional GAN (Flower Generation)

Designed and trained a Conditional GAN to generate realistic flower images from textual descriptions. This project explored the fusion of computer vision and text-to-image generative modeling.

  • Used the Oxford-102 Flowers dataset with pre-computed embeddings.
GANs
Text-to-Image
Deep Learning
Generative AI
PyTorch
Credit Risk Predictor

Classified loan applicants based on risk levels by EDA, feature engineering, and trained an XGBoost ensemble model.

  • Achieved 93.60% mAE and an average mAP of 89.56%.
Matplotlib
Sklearn
Numpy
Pandas
Python

Technical Skills

Psst... hovering over the buttons gives more info.

Education

Indraprastha Institute of Information Technology Delhi (IIIT-Delhi)

Bachelor of Technology in Computer Science and Applied Mathematics

2021 - 2025

Delhi, India

  • Google Developer Student Club: ML Engineer
  • Entrepreneurship Cell: Web Developer
  • Sponsorship team for cultural fest at IIITD

Key Coursework:

Artificial Intelligence
Statistical Machine Learning
NLP
Advanced Programming
AI Ethics
DBMS
Complete Coursework ↗

Get In Touch

My inbox is always open. Whether you have a question or just want to say hi, I’ll try my best to get back to you!

Say Hello