Experience

Built Python microservices and LLM applications powering recommendation and insight workflows.

• Engineered scalable microservices on GKE for recommendation pipelines, ETL orchestration, API integrations, and data processing.
• Built an LLM application using RAG, OpenAI APIs, MCP, and tool-calling agents to extract structured insights from raw audio and content streams.

Architected fault-tolerant data systems and AI-powered tooling to support faculty research at scale.

• Architected fault-tolerant distributed ingestion systems processing 10TB+ of raw data and improving overall pipeline performance by ~70%.
• Built robust ingestion and web scraping infrastructure to collect millions of structured and unstructured financial records.
• Developed AI agents using RAG, vector databases, and OpenAI APIs to automate insight extraction from unstructured datasets.

Shipped full-stack features and data tooling across mission-critical applications.

• Built a data pipeline to parse, predict, and generate GeoJSON outputs from active satellite feeds, reducing runtime by ~40%.
• Contributed across multiple full-stack apps end-to-end — improving performance, scalability, and security.
• Identified and resolved critical issues in production deployments, improving stability and reliability for clients.

Research in machine learning and systems security focused on threat detection.

• Worked on threat detection and SIEM rulesets for Windows, generating datasets via SOCBED and AMIDES environments.
• Evaluated research-grade systems for ML-driven threat detection, applying security principles end-to-end.
• Replicated multiple papers via reverse engineering, demonstrating limitations of generic SIGMA rulesets.

First engineer of the Karpuragaurai Development Wing — owned core platform development and the initial technology roadmap.

• Architected and shipped scalable cryptocurrency forensics and disk forensics applications from a blank slate.
• Built UI and REST APIs in C++ kernels and Python, improving software performance by ~70% via distributed processing.
• Developed AI-driven forensic agents to automate evidence analysis, anomaly detection, and investigation workflows.
• Led a product development team to optimize algorithms and architecture for diverse client deployments.
• Streamlined deployment and testing automation by implementing DevOps, optimizing cloud and on-prem container infrastructure.

Hands-on training internship deploying end-to-end solutions on AWS.

• Completed a structured training program building production-grade solutions on AWS Cloud.
• Deployed services using CloudFront, Amplify, IAM, WAF, EC2, CI/CD, Lambda, S3, RDS, and DynamoDB.