Mustansir Kapasi

Software Engineer

Neosoft•  June 2024 - August 2024•  Wilmington, DE

• Engineered a distributed, scalable user profile management system, supporting 10,000+ concurrent requests and reducing admin time by 30%. • Bolstered security by implementing robust user authentication, authorization, production-grade A/B and unit testing, reducing security incidents by 25%. • Elevated user experience by integrating Google Sign-In and a user-friendly interface, increasing user satisfaction by 15% (user surveys). • Accelerated data retrieval by implementing pagination and search functionality, reducing data retrieval time by 40% for admins. • Partnered with cross-functional teams, designers, and senior developers to transform wireframes into responsive, functional web interfaces, ensuring seamless experiences across devices.

Software Engineer

Infosys•  October 2022 - September 2023•  Bengaluru, Karnataka

Designed and deployed a scalable and robust product configuration solution with digital catalogues in Node.js and Agile, enhancing customer search capabilities based on product features which sped up the sales cycle by 30%, enabling precise product configuration and cost estimation for Johnson Controls-Hitachi Air. Minimised 25% of latency by configuring the applications thread pool, thereby preventing bottlenecks for T-Mobile. Worked in the backend using Node.js to make APIs in different T-Mobile microservices and scaled the distributed system by writing industry standard clean code. Streamlined the frontend with React.js and Angular to build reusable UI components for the T-Mobile app. Worked with Infosys Equinox Gen AI team where I performed fine-tuning on the large language model (LLM) GPT and Llama for sentiment analysis, and summarisation of customer reviews thus rectifying the generation of the FAQ section by 60.2 % for the Internal Infosys RAG e-commerce AI bot using Langchain, huggingface, Bart and Pytorch.

Software Development Intern

NeuralTechSoft Pvt.Ltd•  February 2021 - July 2021•  Mumbai, Maharashtra

Built a react native app which scraped the web for daily financial news for fundamental analysis. Its features included the prediction of live share prices for futures trading using the LSTM machine learning model in Python, Vader, Beautiful Soup and Crontab which gave the company a profit of 100 $ per day on NASDAQ. Created a regression model for the client app Mpower-U to predict loan recovery and made reusable UI components in React, APIs using NodeJs, ExpressJS, MongoDB and customer gRPC endpoints for a client MPower-U which increased 70 % more customers for that client.