Charith Purushotham

My Portfolio

Showcasing innovative projects in Machine Learning, Deep Learning, and Software Development

7+
Projects
4+
Achievements
3+
Tech Stacks

Deep Learning & OpenCV Projects

5 innovative AI/ML projects

Reinforcement Learning-Based Dialogue Summarization

Developed a lightweight LSTM-based attention model for dialogue summarization on the SAMSum dataset, integrating a modular RL wrapper (SCST, PPO) that improved BLEU by +5.42% over baseline without pretrained transformers.
PyTorch RL (SCST, PPO) LSTM NLP
View Paper View GitHub

Multimodal Zero-Shot Learning for Unseen Concepts

Designed a knowledge graph-based GCN model integrating CLIP text embeddings with semantic predicate-level relationships to classify images without labeled data. Incorporated attention-based edge features and custom prototype refinement loss.
PyTorch CLIP GCN Knowledge Graphs
View Paper View GitHub

Human Pose Estimation in Fitness Tracking and Guidance

AI-based tool for real-time exercise monitoring using BlazePose, Yoga pose classification, and automated data reporting.
OpenCV Deep Learning MediaPipe Node.js HTML CSS
Research Paper View Project

Contour Tracking and Detection of Heart Chamber

Developed methods for detecting blood chambers in echo video frames using U-Net segmentation technique in a hackathon.
OpenCV Deep Learning (DeepGaze) PyDICOM Python
Research Paper View Project

RePostureAI

An AI-powered app that corrects posture in real-time, tracks reps, and ensures safe, effective workouts for injury-free fitness.
Python OpenCV Mediapipe Streamlit Gemini AI
View GitHub

Machine Learning & EDA Projects

3 data analysis projects

LinkedIn Job Market Analysis and Recommendation System

Developed a forecasting tool using LinkedIn data and ML models (ARIMA, clustering) to uncover job trends and salary insights.
Python Machine Learning Scikit-learn Matplotlib Seaborn
View GitHub

Document Classification System in Apache Spark for Big Data

Developed a document classification system using PySpark to handle big data and optimize document processing for multi-class classification.
PySpark Article API Machine Learning
View GitHub

Format Matters: An Empirical Study of Data Storage Format Impact on ML Training Pipelines

Benchmarked data formats across CIFAR-10 image training and 1M-row tabular ML workloads, quantifying impact on I/O throughput, load time, and storage efficiency while showing training accuracy remained format-neutral.
Python PyTorch Systems for ML I/O Optimization Benchmarking
View Paper View GitHub

Software Development Projects

2 full-stack applications

Chat Application

A simple chat application using React Chat Engine with real-time messaging and authentication features.
React.js React Chat Engine
View Project

E-Commerce Website

A full-featured e-commerce platform with customer features (browse, cart, secure checkout with Stripe, order history) and admin dashboard (sales analytics, product/order/user management). Built with Next.js 15, TypeScript, Prisma, and NextAuth.js authentication.
Next.js 15 TypeScript Prisma ORM NextAuth.js Stripe Tailwind CSS shadcn/ui SQLite
View Project

Achievements & Publications

Recognition and awards

Research Publication: Nuclear Fuel Characterization Using Machine Learning

Co-authored "Bridging Multimodal Microscopy for Advanced Characterization on Nuclear Fuel Using Machine Learning", published in Frontiers in Mechanical Engineering – Digital Manufacturing, Vol. 11, 2025. Demonstrated a transfer-learning deep-learning framework outperforming four state-of-the-art models for cross-scale defect segmentation in nuclear materials.

3rd Place, T9-MediHack 2025 (24hr Hackathon)

Awarded for "RePosture AI" project in the AI track. A tool for posture correction using computer vision and real-time feedback.

7th Place, AWS Jam Hackathon - CU Boulder

Placed 7th out of 20 teams, solving 9/13 security challenges using AWS services, including Lambda, CloudTrail, DynamoDB, S3, VPC, IAM, and SageMaker.

Best Project of the Year (Computer Science Category)

Awarded for the project "Human pose estimation in fitness tracking and guidance" at VTU, Belagavi, August 2022, supported by KSCST.

Extracurricular Activities

Leadership and community involvement

Teaching Assistant, Distributed Systems - CU Boulder

Jan 2026 - Present

Support Prof. Mark Zhao in running the course by holding office hours, assisting with assignment/project design, grading, and helping students debug distributed systems concepts and implementations.

Technical/Data Consultant, Data Buffs

Jan 2025 - Present

Working on data extraction, backend development, and statistical analysis to provide data-driven insights.

Co-Founder, Data Science and AI-ML Student Club (Layers)

Dec 2021 - Present

Lead Layers Club to promote knowledge sharing, networking, and skill development in Data Science and AI-ML.