Apr 2024 – May 2025 · 1 yr 1 mo · Bengaluru, Karnataka, India · Hybrid

• Building an AI platform where non-technical team members can build and deploy custom internal tools, such as AI Apps, CRM or ERP systems, using prompts.
• Fine-tuned open-source models using ASTs to achieve 100% accuracy in internal tool code generation.
• Designed a JSON-based design system language enabling LLMs to generate dynamic UI components.
• Led backend, frontend, and CI/CD pipeline (CircleCI) with full deployment automation using Kuber-
• netes.
• Built an AI agent orchestration engine for complex workflows via OpenAI tool chaining with memory.
• Delivered a no-code AI platform for non-technical users to build agents.
• Worked on RAG, model finetuning, AI tool calling, observability while building AI agents

May 2022 – Mar 2024 · 1 yr 10 mos · Bengaluru, Karnataka, India

• Built India's First Credit Sharing platform at Credwise. Credwise is a platform that allows users to bring their entire family expenses on a single credit card resulting in maximum savings for customers.
• Scaled Credwise to 150k users across Android and IOS devices. Credwise was acquired in Dec 23.
• Led a team of 10+ members across Growth, Product, Marketing, and Technology.
• We raised our pre-seed cheque from 100x.vc
• Led end-to-end software lifecycle: PRDs, HL/LL design, agile sprints, staging, production.
• Integrated with Visa, Cashfree, Razorpay for prepaid card offering of Credwise; scaled to 150k+ MAU.
• Built a high-observability stack with New Relic, Datadog, Mixpanel, GA.
• Enforced 85%+ code coverage, CI/CD workflows, and unit testing using CircleCI.
• Directly managed a team of 8 developers.
• Stack: Java, Python, Flutter, PostgreSQL, MongoDB, Redis, CircleCI, Kubernetes, AWS

Oct 2020 – May 2022 · 1 yr 7 mos

• Developed IAM authorization stack used by hundreds of microservices across Nutanix clusters.
• Redeveloped monolithic IAM codebase into microservice architecure using Golang.
• Rolled out 30+ SDKs for multi-language client integrations.
• Improved P99 latency to under 10ms for authorization endpoints.
• Tech stack: Go, Python, Java, Kubernetes, Postgres, Redis, OData, ETags.

Dec 2018 – Sep 2020 · 1 yr 9 mos · Bangalore

• Sizing : Worked on a highly scalable size recommendation service. The system handles 200,000 rpm peak load. Restructured the entire redis cluster and servers in order to handle the peak load. Used multiprocessing/fast dumping processes (along with MR jobs) to enable faster dump of data in redis and further reduce the service downtime while doing the model refresh . Added features like AB test, redesigned the code for parallel AB tests.
• Footwear/Apparel recommendations : Worked on api contracts design, development, test- ing of entire footwear and apparel PS0 recommendations thus enabling recommendations for first time users. Apart from backend, also worked on graphql and golang to do gateway changes for this feature.
• Cross-sell: Worked on redis, gunicorn with flask for running parallel processes and handling the traffic. The response time of the DS models was very high (300ms) and hence certain parts of data was kept in-memory in order to reduce the number of db calls/latency.
• Data-pipeline: Managing clickstream and visitstream data pipeline.

Jun 2018 – Nov 2018 · 5 mos · Gurgaon

• Worked on predicting affinity of a user to be spending in car rental and airline industry and further their spend propensity in those sectors. Used GBM for modelling.

Jul 2017 – May 2018 · 10 mos

• 1) Mining interesting patterns in airline data and further analyzing how the delay of one flight affect the Schedule Block Time (SBT) of other flights.
• 2) Analyzing dense subgraphs of airline data at different time stamps and further optimize the schedule block time of airlines.
• 3) Worked on a new approach to divide the graph in communities based on the maximum entropy theorem and developed a new greedy hill-climber to achieve it.