May 2025 – Present · 10 mos · Bengaluru, Karnataka, India · Hybrid

• Helping early-stage startups and teams integrate AI into their products and workflows.
• I work across AI strategy, backend infrastructure, and data systems, from idea to prototype. My focus is on solving hard technical problems with product sensitivity, and creating solutions that deliver real business value.
• Available for short-term consulting, long-term collaborations, or sparring on technical direction with founders.

Jun 2024 – May 2025 · 11 mos · Bengaluru, Karnataka, India · Hybrid

• Collaborated closely with Uber’s Heavy Commercial Vehicle (HCV) Engineering and Product teams to deliver end-to-end data solutions, driving significant business impact.
• Architected and implemented a comprehensive billing data pipeline for the HCV product, reducing billing quote turnaround from weeks to days by optimizing asynchronous distributed backend jobs and managing complex semi-structured data workflows.
• Developed advanced data solutions for HCV scheduling datasets, empowering Operations teams with deeper insights and streamlining their workflows to save time and effort.
• Revamped real-time GPS data processing for HCV vehicle arrival/departure estimation by identifying critical bugs and designing a robust, edge-case resilient solution; improved data quality by reducing null value rates from over 30% to under 10%.
• The GPS solution enhanced ML model accuracy for arrival/departure predictions and fraud detection by integrating enriched datasets through collaboration with the data engineering fulfillment team.
• Partnered with the Uber Registra team to build real-time fleet-level metrics pipelines for fleet partners, improving on senior developer proposals and delivering a fully optimized production pipeline.
• Key Technologies: Apache Spark, GoLang, TeraBlob (Uber’s distributed blob storage), Piper (Airflow alternative), Apache Pinot

Oct 2023 – Jun 2024 · 8 mos · Bengaluru, Karnataka, India

• Spearheaded the enhancement of the core Ads Relevance service, gaining deep expertise in optimizing ad delivery and performance at scale.
• Designed and developed a comprehensive feature health monitoring framework, encompassing anomaly detection, comprehensive dashboards, and automated alerts.
• Mastered the intricacies of 200+ features to build an intuitive system enabling even non-technical team members to effortlessly add and monitor feature checks.
• Proactively identified and resolved critical bugs through the health check framework, significantly improving the accuracy and effectiveness of the Ads ranking algorithm.
• Prevented major production issues by catching elusive bugs early, safeguarding service reliability and driving consistent ad performance improvements.

Apr 2023 – Oct 2023 · 6 mos · Bengaluru, Karnataka, India

• Built the org’s first user-post A/B testing framework to target post-level cold start and late lifecycle engagement; designed end-to-end from hashing strategy to candidate generation, accelerating feedback loops and boosting user engagement across the content lifecycle. This was later used by other teams to run more such experiments.
• Proposed and developed a genetic algorithm–based ranker tuning system to optimize non-differentiable business objectives; ran A/B tests focused on time spent, delivering measurable lifts across core engagement metrics.

Feb 2022 – Mar 2023 · 1 yr 1 mo · Bengaluru, Karnataka, India

• Optimized ShareChat's core RecSys service, reducing latency by 30% and improving feed load time by 25% through deep profiling and targeted improvements in concurrency, caching, and serialization efficiency.
• Designed and deployed a two-tower Candidate Generator(RecSys Algo System) end-to-end—including feature pipelines—as part of the content lifecycle team’s candidate generation initiative.
• Scaled candidate generation capacity 5x (from 100K to 500K posts), unlocking a 25% boost in user engagement and retention gains; led everything from model deployment and memory optimization to A/B testing and hypothesis validation.
• Built and launched heuristic-based candidate generators for festival-specific content; derived insights that challenged org-wide assumptions and transferred operations to growth teams.
• Introduced a parallel A/B testing framework for experimental RecSys workflows—later adopted across the org to enable faster iteration cycles.

Jul 2021 – Feb 2022 · 7 mos · Bangalore Urban, Karnataka, India

• Refactored code of the data warehouse platform to improve maintainability.
• Pitched and helped build a natural language–to–SQL tool during a team hackathon—enabling non-technical users to query data before mainstream LLMs existed; won first place.
• Conducted in-depth resource usage analysis; proposed and implemented a pooling strategy projected to save ~$400K annually in compute costs.
• Developed dashboards for monitoring and analyzing infrastructure usage, improving visibility and operational decision-making.

Mar 2021 – Jun 2021 · 3 mos · Bengaluru, Karnataka, India

• Collaborated with the RecSys AI team to build production-grade microservices powering large-scale machine learning systems.
• Partnered with PMs to design and launch A/B tests, accelerating experimentation cycles to improve user retention.
• Re-architected a bloated feature-serving system into a unified, formula-driven service—cutting 20+ endpoints down to one and enabling faster, safer deployments.
• Led optimization of the core recommendation service, introducing a caching layer that reduced CPU usage by 20% and became a performance benchmark across teams.
• Technologies:
• Golang, Airflow, Python, Kubernetes, Pub/Sub, BigQuery, Dataflow, JavaScript, Load Testing, Microservices, A/B Testing

Jun 2020 – Aug 2020 · 2 mos · Gurgaon, Haryana, India