Jan 2025 – Jul 2025 · 6 mos · Bengaluru, Karnataka, India · On-site

• Deployed Prometheus to monitor system performance metrics such as CPU usage, memory consumption, and API response times, along with setting up real-time alerts using Prometheus Alertmanager, and reduced incident response time.
• Developed an ACL system with role-based access controls and granular permission levels to manage user roles and permissions securely.
• Upgraded the codebase to GHC 9.7 (Haskell compiler) using Nix Flakes, improving dependency management and reproducibility.
• Reduced build time by 25% through automated environment setup, Docker integration, and process-compose orchestration for Elasticsearch and Postgres.
• Build a compiler and configure it for multiple language support

Aug 2023 – Dec 2024 · 1 yr 4 mos · Bengaluru, Karnataka, India · On-site

• Designed the backend architecture from scratch
• Implemented asynchronous logging and raw SQL query logging, reducing debugging time by 50%.
• Improved server reliability and scalability by handling 10,000+ concurrent requests with minimal downtime.
• Designed and implemented Elasticsearch-based APIs to handle millions of data records with sub-second latency using semantic vector models and KNN similarity search.

Jan 2023 – Jul 2023 · 6 mos · Bengaluru, Karnataka, India · On-site

• Developed and launched a Driver Referral Program, achieving a 35% increase in cost-effective customer acquisition through trust and recommendations.
• Implemented a Referral Reward System, leading to a 40% boost in driver recruitment within the first month of launch.
• Increased driver engagement metrics by y automating milestone-based rewards, improving retention rates for referred drivers.
• Integrated Google NearbySearch API as a fallback for Geocode API, retaining 3% of users whose locations couldn’t be fetched by drivers.
• Reduced ride drop-offs by 10% in scenarios where geolocation errors occurred, improving the overall ride completion rate.
• Reduced OPS team intervention time by 40%, enabling them to manage ambiguous ride statuses more efficiently.
• Migrated ride scheduling logic from Postgres (PSQL) to Redis, reducing API latency by 40%.
• Lowered infrastructure costs by 25%, optimising backend performance during high-traffic hours.
• Improved average response time from 300ms to 180ms, enhancing the user experience for both drivers and customers.

Apr 2022 – May 2022 · 1 mo · Cape Town, Western Cape, South Africa

• Zaio is dedicated to creating training programs that promote lifelong learning in order to cultivate leaders that will drive the future of African and Global innovation.

Feb 2022 – Mar 2022 · 1 mo

Nov 2021 – Dec 2021 · 1 mo

Sep 2021 – Nov 2021 · 2 mos

Jul 2021 – Aug 2021 · 1 mo

• Contributor for DCP 2021.

Jun 2021 – Aug 2021 · 2 mos

Jun 2021 – Jul 2021 · 1 mo · Bengaluru, Karnataka, India