Working on improving performance of RAG chatbot in Cisco XDR with LangChain and Pytorch.

• Developed functionality to ingest/process data from 11+ sources using C# and Entity Framework Core, and leveraged Angular to display factory metrics in UI; Reduced existing reporting workflow from 8 hours to seconds.
• Developed key features on large scale machine-learning application that returns oil and gas production forecasts within minutes.
• Wrote end-to-end API’s using FastAPI to generate business-unit-specific production forecast plots from machine learning models, visualized with Plotly’s Python library, and integrated with a front-end built using Angular.• Developed functionality to ingest/process data from 11+ sources using C# and Entity Framework Core, and leveraged Angular to display factory metrics in UI; Reduced existing reporting workflow from 8 hours to seconds.

• Developed key features on large scale machine-learning application that returns oil and gas production forecasts within minutes.

• Wrote end-to-end API’s using FastAPI to generate business-unit-specific production forecast plots from machine learning models, visualized with Plotly’s Python library, and integrated with a front-end built using Angular.

• Automated telecommunication inspections with React Native mobile app and image classification models, bringing in significant revenue by securing lucrative contract with vendor.
• Developed ’Talk to Your Data’ web application leveraging OpenAI’s GPT-4 model and LangChain, in turn mitigating difficulties with data storage across multiple platforms.

Revamped company’s drug sequencing system using Flask back-end and React front-end, enabling seamless selection of multiple data sets simultaneously, resulting in a 50% reduction in drug sequencing time.

• Developed a web-based visualization tool using Flask, Plotly Python, and React to interpret protein structures, significantly enhancing research accuracy.

• Created data analysis software using OpenCV, Pandas, OpenPyxl, and Matplotlib for major client, allowing efficient analysis of 4,000+ Piping and Instrumentation Diagrams.