Jay Shah
Development
Maharashtra, India
Skills
Machine Learning (ML)
About
GitHub JS's skills align with Programmers (Information and Communication Technology). GitHub also has skills associated with IT R&D Professionals (Information and Communication Technology). GitHub JS appears to be an entry-level candidate, with 14 months of experience.
View more
Work Experience
OPENAI_API_Key, FastAPI, TypeScript
March 2024 - April 2024
- Developed a chatbot using LangChain, OPENAI_API_Key Implemented backend with FastAPI, frontend with TypeScript, and MongoDB for storing files. • Technologies: LangChain, OPENAI_API_Key, FastAPI, TypeScript, MongoDB.
Pro
DIALOGUE SUMMARIZATION WITH MICROSOFT PHI-2 LLM
February 2024 - March 2024
- • Fine-tuned the Microsoft Phi-2 Language Model for dialogue summarization task utilizing QLoRA and PEFT techniques to enhance performance. Used Dataset from Hugging Face. • Technologies: PyTorch, Hugging Face, QLoRA, PEFT, Dialogue Datasets.
TEXT2IMAGE SEARCH
January 2024 - February 2024
- • Scraped images and generated embeddings using CLIP for both images and text. Stored embeddings in Qdrant for efficient retrieval. • Technologies Used: Semantic search, CLIP, Vector DB, Web Scraping.
SOFTWARE DEVELOPMENT ENGINEER
SHOPSE
January 2023 - July 2023
- Enhanced refund tracking for Byju's by creating refund timeline visibility using ReactJS on the frontend and implementing backend Flask API's. • Revamped EMI Conversion File Tab for better functionality in Java. • Developed 10+ Python Flask API's in the Ops Panel, automating manual tasks and significantly reducing workload by 50%. • Developed a Java file for seamless SFTP uploads for Bank of Baroda. • Implemented Slack alerts for real-time issue notifications in EMI Conversion Tab. • Resolved production issues efficiently and facilitated timely and budgetfriendly software releases using JIRA
MACHINE LEARNING ENGINEER
LEEWAYHERTZ
September 2023 - Present
- ? Remote • Developed chatbot for Rackspace client using FastAPIs for LLM interaction, employing Google's Vertex AI for embeddings and semantic matching to create chatbots. Utilized Nvidia's NeMo guardrails for response validation. • Constructed 10+ Retrieval Augmented Generation (RAG) systems from scratch for different clients, covering data collection, indexing, retrieval embedding, and semantic matching for optimal performance. • Engineered a document classifier app for client use, enabling classification of uploaded PDFs into different entities. • Created 10+ Customer Support bots using Microsoft Autogen for different clients, integrating RAG for enhanced responses and user interaction in group chats. • Performed Data Engineering in Google BigQuery to make data compatible for Google Recommendation AI, and also trained recommendation models for "Recommended for you " and "Similar items " functionalities.
Education
