About Me
Apratim Mishra
PhD Candidate in Information Sciences
University of Illinois at Urbana-Champaign
About Me
Hello! I am a PhD candidate in Information Sciences at the University of Illinois at Urbana-Champaign. My research focuses on machine learning and natural language processing (NLP) applications. I’m passionate about developing innovative solutions to complex problems using advanced AI techniques.
Research Interests
- Natural Language Processing
- Large Language Models
- Graph Neural Networks
- Machine Learning Systems
- MLOps & Cloud Computing
Technical Skills
- Core ML: PyTorch, TensorFlow, Transformers, LangChain, RAG
- Data Science: Python (pandas, NumPy, scikit-learn, nltk), R, SQL
- MLOps: W&B, Comet, ZenML, Ray, DeepSpeed, Accelerate
- Cloud & Big Data: AWS, GCP, Apache Spark, Hive, Snowflake
- DevOps: Docker, Git, Linux, Kubeflow, Hopsworks
Achievements
- Published research in top conferences
- Optimized ML pipelines reducing costs by 15%
- Developed production-grade NLP systems
- Implemented scalable ML infrastructure
Projects
Professional Projects
Protein Language Models
Utilized protein language models (PLMs) and graph neural networks (GNNs) to derive insights and predict complex biological behaviors. Implemented optimized training pipelines with PyTorch for large-scale protein sequence analysis.
Learn More →NLP Pipeline for Quotation Extraction
Engineered an NLP pipeline for quotation extraction and entity classification leveraging tools like spaCy and Stanford CoreNLP. The system achieved 87% accuracy on complex news article datasets.
Learn More →Energy Load Prediction
Analyzed energy load trends using Python by employing models like ARIMA, XGBoost, LightGBM, and LSTMs. Developed a predictive system that forecasts energy consumption patterns with 92% accuracy.
Learn More →Personal Projects
Yelp Restaurant Recommendations with GNNs
A Graph Neural Network (GNN) based recommendation system for restaurants using the Yelp dataset, implementing heterogeneous graph models for personalized recommendations.
Notion RAG
A Retrieval-Augmented Generation (RAG) system that leverages Notion data for AI-powered question answering and knowledge retrieval, with a Flask API and vector search integration.
H&M Fashion Recommendation System
A machine learning recommendation system for H&M products using a hybrid approach combining collaborative filtering and content-based methods for personalized fashion recommendations.
Audio App
A Next.js application for audio processing and playback with modern UI and advanced features. Implements Web Audio API for real-time audio manipulation.
Expo v1
An open-source framework for making universal native apps with React that runs on Android, iOS, and the web. Features cross-platform compatibility.
ScenicBayMapper
A mapping application for discovering and navigating to scenic locations around the Bay Area. Integrates with Google Maps API and features location-based recommendations.
Recent Publications
For a complete list of my publications and citations, please visit my Google Scholar profile.