Astha Aggarwal

About Astha Aggarwal

Current experience: Senior software developer in Goldman Sachs (Core engineering team).Past Experience: Backend developer at optim global solutions(UHG)I had internship at Dentalkart where I worked on Django and developed two projects. Key areas: web development, api developmentHave good command in data structures, algorithms and problem solving.Gold medalist in my branch with an overall CGPI - 9.38

Experience

1. Software Engineer Ii(L-62)

   Microsoft

   Jun 2024 — Present · Bengaluru, IN

   As part of the Financial Services team within Microsoft Cloud for Industry, I am working on a range of AI-driven and DevOps-focused projects designed to improve productivity, reduce manual effort, and deliver intelligent solutions to enterprise customers. Key contributions include- AI Copilots1. Financial News Copilot:Designed and developed an AI-based Copilot that summarizes and presents real-time financial news to users. It allows natural language queries and returns structured, relevant insights — improving decision-making speed for financial analysts.2. Domain-specific Copilot with RAG:Created a specialized Copilot by:a. Scraping structured and unstructured data from trusted financial websites.b. Integrating a Retrieval-Augmented Generation (RAG) pipeline that combines LLM reasoning with custom dataset search, ensuring high accuracy and relevancy of answers.c. Supporting semantic search and query refinement for better user experience- DevOps & Automation1. Azure DevOps Pipelines:Developed and maintained YAML pipelines to automate:a. Solution migration across environments (e.g, dev → test → prod).b. Pull request validation (including static checks).c. Artifact packaging and deployment.d. These pipelines improved reliability and reduced deployment time.2. User and Environment Migration Automation:Authored PowerShell scripts to:a. Extract and migrate users, roles, environment variables, and metadata from one tenant to another.b. Enable automated failover in case of tenant expiry.This replaced error-prone manual work and ensured business continuity- LLM Evaluation & Analyticsa. Built an LLM evaluation tool that:b. Parses chat transcripts from AI copilots.c. Calculates quality metrics (e.g. helpfulness, accuracy, precision, summarization) based on use case.d. Outputs results to a Power BI dashboard for easy review by product and engineering teams.This enabled faster tuning of models and better visibility into Copilot performance.

Education

• National Institute of Technology Hamirpur-Alumni

  Bachelor of Technology, Electronics and communication

  2016 — 2020

• Modern Public School, Bhiwadi

  PCM