Jun 2024 – Jan 2026 · 1 yr 7 mos · Bengaluru, Karnataka, India · On-site

• Owned and scaled the CX Incident Management System (IMS) — a multi-tenant, mission-critical reliability platform serving 180M+ customers, 10K+ command center agents, Sellers, Logistics (EKart), and partner businesses including Myntra, Shopsy, and Jeeves. Processed ~2M incident creations/day with high durability and isolation.
• Led a 17-engineer org with 0% FTE attrition over ~1.5 years, enabling 2 formal career growth moves/promotions, reflecting a high-trust, psychologically safe, and ownership-driven team culture.
• Built and shipped from scratch a generic low-code workflow orchestration platform, accelerating go-to-market (GTM) by ~50% for CX automation use cases.
• Launched Chatbot-driven automated Return Creation using the new platform, reducing customer contacts by 22 bps and improving user experience at scale.
• Drove production readiness for Big Billion Days (BBD) 2024 & 2025, scaling IMS to handle ~20K incident updates/sec under peak load with P99 latency ~600 ms, ensuring real-time reliability for critical CX operations.

May 2023 – Jun 2024 · 1 yr 1 mo

• Single-threaded owner of the Manual Annotations platform, powering ~5M image and ~2M video annotations/month for continuous ML model training and evaluation.
• Single-threaded owner of a security-compliant offline ML batch processing platform, used by 15+ internal teams (across 4 timezones), orchestrating ~60K batch jobs/month for data prep, model evaluation, and experimentation.
• Directly managing 7 SDEs, driving 85%+ retention, fostering high ownership, mentorship, and career growth.

Jul 2021 – Apr 2023 · 1 yr 9 mos

• Built a Kibana-based near real-time (vs earlier 24-hour cron-based) dashboard to provide transparency of manual annotations (labeling tasks) throughput and backlog.
• 1-click submission of manual annotations via workflows, reducing TAT from ~8 hours to <2 hours.

Aug 2020 – Jul 2021 · 11 mos

• - Designed and delivered rule rule-based logistics fee framework (with auditing capability) based on Geo, client (Buyer/Seller), and business verticals. This initiative led to an increase in logistics fee revenue by ~30%.

Dec 2017 – Aug 2020 · 2 yrs 8 mos · Bengaluru, Karnataka, India

• Built Conversational Search 2.0 from scratch with single-threaded ownership of the HLD and the LLD of the entire pipeline. Enhanced coverage by ~4X and CTR by ~40%.
• Non-Functional Requirements (NFR) benchmarking/optimisations for Flipkart Search Federator for Big Billion Days 2018. Scaled Search Federator to ~100K QPS with P99 latency of ~550 ms.
• Non-Functional Requirements (NFR) benchmarking/optimisations for Flipkart Conversational Search for Big Billion Days 2018. Scaled Conversational Search to ~65K QPS with P99 latency of ~55 ms. Analysed thread context switches in Hystrix pools and fine-tuned the thread allocations.

Feb 2017 – Dec 2017 · 10 mos

• Design and implementation of a new (C++ based) stack for TurboTax Desktop. This new stack saves the heavy KTLO efforts on the legacy stack, and enhances TTM agility by 50%.
• Integration of the above stack with TurboTax Windows (C#) and TurboTax Mac (Objective-C).

Aug 2015 – Jan 2017 · 1 yr 5 mos

• Conducting an A/B experiment on ~5 million TurboTax customers on promotional offers. The aim of the experiment was to infer the best offer, to enhance customer retention, and new customer acquisition.
• Enhancing the TurboTax installer success ratio by fixing filesystem permissions-related issues. Resulted in the reduction of ~4K customer care calls per year.

Jul 2014 – Dec 2014 · 5 mos · Bengaluru, Karnataka, India

• POC on different graph databases like Neo4J, OrientDB, and Titan+HBase on the basis of
• import and traversal time using Tinkerpop APIs.
• Profiling and Logging using AspectJ.