I work on building AI-powered applications, from figuring out the right approach to implementing, testing, and deploying them. My focus is on making things work reliably beyond the initial prototype. At Zion Cloud Solutions, I work on applied AI engineering, building and improving AI-powered search experiences and generative AI solutions. My work spans search relevance, LLM evaluation, data analysis, system optimization, and production engineering. I care about measurable improvements, rigorous testing, and solving the underlying problem rather than just making a demo work. I also build AI agents using Python and Google's Agent Development Kit (ADK), translating project requirements into practical agent workflows and deployable solutions. Outside client projects, I developed Visualizer Agent, an AI-powered data analyst that turns natural-language questions into charts and dashboards over company databases. I'm building it with a focus on the complete product experience, including authentication, data access controls, and agent guardrails. I also won 1st place at the Nasiko Open-Source Build-a-thon for an optimization to an LLM router, with my contribution merged upstream. My background in mechanical and R&D engineering shaped how I approach software today: measure results, validate assumptions, and let evidence guide engineering decisions. Tech I work with: Python, FastAPI, LLMs, AI agents, Google ADK, Gemini, Vertex AI, BigQuery, Firestore, Cloud SQL, Cloud Run, Rust, and CI/CD. Certifications: Google Professional Machine Learning Engineer · Google Associate Cloud Engineer · Databricks Data Engineer Associate Always happy to connect with people building AI systems, experimenting with new ideas, and taking projects from prototype to production.
AI Engineer
Zion Cloud Solutions Private Limited
- Optimized AI-powered product search, reducing latency by ~6× (1.6s to 0.26s) and increasing query coverage by 17% by consolidating three separate search indexes into a unified, ranked search layer using Vertex AI Search, Python, and FastAPI. The team adopted the new approach in place of the previous pipeline.
- Delivered five releases of a product compatibility feature that matches customer equipment to compatible parts. Analyzed 64K product records to refine matching logic and request additional customer details only when needed to determine the right match.
- Built a product search API and supporting admin panel for part-number and keyword searches, with autosuggest and live pricing and inventory lookups, backed by 500+ automated tests.
- Contributed to AI-powered client solutions by building intelligent agents with Python and Google's Agent Development Kit (ADK), designing agent workflows and implementing generative AI capabilities to meet project requirements.
R&D Engineer
Pavakah Energy Private Limited (IIT Hyderabad)
- Engineered automated data-acquisition pipelines and multi-sensor setups, reducing raw time-series processing time by 60% during long-duration qualification tests.
- Applied Design of Experiments (DoE) and statistical machine learning to bench-validation data, reducing physical test iterations by 40% and improving system efficiency by approximately 12%.
- Developed simulation models using MATLAB, FreeCAD, and OpenFOAM to guide data-informed prototype iterations from CAD through fabrication.
- Applied generative AI workflows to investigate R&D bottlenecks and support faster experimental development cycles.
Technical Sales Engineer (CAE)
CADFEM APAC Private Limited
- Consulted 10+ cross-functional engineering clients on computational modeling and data-driven design decisions.
- Delivered technical presentations on structural and thermal simulation solutions; supported prospect qualification and proposal development with cost framing.
Manufacturing Engineering Intern
Tata Lockheed Martin Aerostructures Limited
- Managed engineering data in SAP, including Bills of Materials (BOMs), routings, and assembly instructions.
- Standardized technical documentation and coordinated planning updates with production teams.
Nasiko Build-A-Thon — 1st Prize | Compact Tool Schemas
• Earned 1st prize for an open-source Rust contribution improving LLM tool-calling efficiency. • Built compact tool-schema encoding/decoding, strict validation, streaming decoding, and opt-in router support; reported 42.7% token reduction on the public sample and 83% vs. 64% correct calls across 5 models / 21 cases.
Visualizer Agent
• Built an AI agent that lets non-technical users ask questions in plain English and get charts and dashboards from company databases. • Added secure sign-in with Google accounts, per-user database permissions and live BigQuery and Cloud SQL connections, so each user sees only the data they are allowed to.
B.E., Mechanical Engineering
Vasavi College of Engineering
Databricks Certified Data Engineer Associate
Databricks
Oct 2026Associate Cloud Engineer Certification
Professional Machine Learning Engineer Certification
Attendee · Oct 24, 2026