Feb 2025 – Jun 2025 · 4 mos · Gurgaon · On-site

Jan 2025 – Feb 2025 · 1 mo · Bengaluru, Karnataka, India · On-site

• Migrated Even Steven from Python to Go, decreasing response time by 20% and eliminating intra network calls to the python service.
• Created a Langfuse Go Client for establishing trace for llm requests, hereby preventing external dependency for the codebase.
• Migrated data from Qdrant to Firestore, which enabled turning off GKE Cluster which hosted Qdrant service
• Created medical summarisation engine for summarising medical history for reachouts.

May 2024 – Jul 2024 · 2 mos · Bengaluru, Karnataka, India · On-site

• Worked on designing, streamlining and implementing a new POC feature in the Google Search App
• Delivered the main project well in time, and worked on an additional project which comprised of work related to making server side changes and learnings related to concepts of microservice based architecture and server side rendering
• Worked with multiple ecosystems and components, gained familiarity with complex code patterns and modern system architectures.

May 2023 – Jul 2023 · 2 mos · Gurugram, Haryana, India · On-site

• Delivered a feasibility study project on AR/VR, streamlining possibilities for the firm to work on Vision Pro (and related) next gen ecosystems.
• Created an inhouse version of an dependency, which was 60% less space consuming and reducing technical debt on the codebase.
• Learned industry standards and translating theory knowledge into practical aspect. Worked on minor bugs and fixes and learned about workflow before pushing in production.

Apr 2023 – May 2023 · 1 mo · Remote

• Worked to make a production ready app for the travel based startup.

Dec 2022 – Jan 2023 · 1 mo · Remote

• Created a local network payment interface for the transactions within an institute.
• Integrated Razorpay within the App to add money to the account, which could be used to transfer money to peers and in shops within the network through the QR generated.
• The ecosystem was compatible with Institute Ids which could be scanned for tap-to-pay kind of functionality.