• Architected horizontally scalable AI services (stateless) and custom worker pools; applied caching, and rate limiting; built Instana operational dashboards (via OTel collector pipeline) to monitor FastAPI latency, pod CPU utilization.

• Built RBAC-enforced agentic AI assistant; developed AI agents in LangGraph using the ReAct pattern

• Implemented eval + safety: hallucination checks, LLM-as-judge (gpt-oss-120b), offline evals(manual + automated); production controls including per-user daily token limits and fallback models for investigation reliability and cost control.

• Led AI agent incident investigation (logs + traces + spans + infra) as Python microservices on Kubernetes; reduced MTTR ~10%.

• Raised Analyst Acceptance Rate by 30%

• As a part of my role I bring Software Engineering solutions to the Cyber Security Domain

• Collaborate and disrupt traditional processes and assess the scope of new solutions to the transformation space

• Implement framework for checking breach of customer data over the internet

Software Architecture Course

● Design and Develop automation framework to centrally trigger Regression Tests of all teams under IBM Information Server (IIS) and Watson Knowledge Catalog (WKC) standalone and CP4D environment.

● Developed and Maintained several Jenkins CI pipelines for Sanity testing of Information server.

● Design custom notifications for CI/CD status as a part of Release Engineering of the IIS components and IBM Watson Knowledge Catalog.

● Reviewed and assisted teams with their continuous integration frameworks.

● Developed RESTful API’s in MEAN stack web application for monitoring test results using ExpressJS, MongoDB and Angular 6 with CRUD functionality and scalable to large number of requests in Agile environment and used by ~10 teams as one stop shop for the results and trends in Regression and Sanity testing for release Engineering.

● Designed and developed performance analysis algorithm for the performance of a self-adaptive system based on response time of the incoming requests for Unmanned Aerial Vehicle (UAV’s).

● Enhanced a novel technique to reduce uncertainty in selection of tactic when making the adaptation

decision.

● Collaborated with senior researcher form Carnegie Melon university

● Defining a trade-off for engineers to monitor the maximum allowed latency of a tactic

● Assurance cases for microservices architecture oriented Autonomous system.

Tools: Azure DevOps, Google Cloud, Azure , Java, Javascript, C++, Web Simulators, Javascript