MS CS @ CU Boulder | swe

Sahasraditya Thyadi  • Peerlist

Feature Extraction and Recognizing Gender of Speaker using Speech signals done as part of an internship at Innovation Lab, BEL.

AI-based Gender Recognition of speech signals

Identification of fourteen common thoracic diseases observed in chest X-rays using deep learning methods.

Classification and Localization of Local Thor

Engineered an intelligent chatbot using Langchain framework and OpenAI's Large Language Model (LLM) for querying cricket statistics Integrated natural language processing to transform user queries into SQL, facilitating rapid retrieval of player statistics from a database containing more than 200,000 entries Incorporated few-shot prompting techniques to enhance the chatbot's understanding of cricket-specific terms and metrics (e.g., strike rate, economy rate, player matchups

Generative AI SQL chat-bot

Developed a full-stack news platform with three main components: Web App: Allows users to register, select interests, and view personalized news. Collector Service: Fetches news from a public API, caches responses using Redis for 30 minutes to reduce API calls. Analyzer Service: Applies NLP (NER via Stanford CoreNLP) to extract keywords and group related articles. Email Service: Sends daily newsletters to users based on their selected topics using RabbitMQ.

News Aggregator

Text2Comic is a web application that utilizes Stable Diffusion to transform ordinary texts into visually captivating comic strips. Designed and developed a scalable full-stack web application using Next.js (frontend) and Flask (backend) to transform text into comic-style visuals, integrating OpenAI GPT-3.5 and Stable Diffusion for narrative processing and comic-style image generation Implemented a cloud-native architecture leveraging Kubernetes for container orchestration, Google Pub/Sub for distributed messaging, Redis for caching, and Google Cloud Storage for output management, ensuring high availability, scalability, and optimized performance.