Jul 2025 – Oct 2025 · 3 mos · Berlin, Germany · Remote

• Transitioning from solving well-defined problems to defining the problems worth solving.
• Built https://audiojot.ai/, a privacy friendly voice diary. Explored how vibe coding and indie tools can help accelerate end to end production grade product development.

Oct 2024 – Jun 2025 · 8 mos · Remote

Nov 2021 – Jan 2026 · 4 yrs 2 mos · Berlin, Germany · Hybrid

• Development and maintenance of software localised in 40+ languages for 1 Million riders globally in 75+ countries
• Splitting monolith and creating microservices in Golang from scratch and helping scale the system
• Improved GDPR compliance by cutting non permissible rider location tracking by 95%, avoiding ~€21M penalty.
• Led MVP architecture enabling faster time-to-market and scalable backend; drove workshops, hackathons, and cross-team
• collaboration to deliver a POC that became a core component of the tribe. Approx 40 stakeholders
• Solved the problem of error translation and showing meaningful translated error messages to riders using a very pragmatic solution which requires minimal changes in all gateway services
• Lead engineer on rider safety initiative which integrates with a telematics solution provider to reduce rider accidents by 70% in Talabat (brand of DH in middle east) operating countries
• Lead engineer on incentive which provides history of orders to vendors as well as rate the vendors
• Managing infrastructure of services via terraform
• Part of oncall team
• Host of backend channel guild comprising of 9 engineers. We reduced median alerts per oncall shift by 62% as well as identify and fix health of our services.
• Created resource localization SDK in JAVA for our JVM services to move towards backend for frontend solution and hence decoupling from release cycles. Addressing this proactively saved our domain 1 quarter for Android, backend and iOS

Jun 2019 – Oct 2021 · 2 yrs 4 mos · Bangalore

• Part of oncall rotation team
• Product focused features:
• Split and stitch feature: Splitting larger data annotation job into smaller pieces to enable faster labelling.
• Submission and rework feature: Centralised data submission process to customer. Earlier it was a done in a decentralised way. This feature also tracks data submission and rework process and helps identify inefficiencies.
• Core tech features:
• Built incident identification micro service from scratch which helpes us identify inconsistencies in data because of non fault tolerant code. Also helps us identify incidents before customers report them and helps us resolve them, hence improving customer experience.
• Large data deserialisation optimisation. I have worked on reducing core microservice memory consumption which helped increase its scalability by 2x
• Dropwizard (java lightweight framework) enhancements to make it transactional
• Refactoring code for optimisation as well as removing dependency from old micro services which are difficult to maintain
• Single Sign On implementation in our system - Aimed at providing enterprise customers easy and controlled access to our SaaS tool.
• Media upload - An intuitive drag and drop feature to help clients quick create jobs for annotation and get started with the product.
• Subscription management and billing system integration with our Saas product
• Core tech features:
• Working on engineering fault tolerant distributed SaaS system
• Worked on Job Cancellation feature which allows end user to stop data labelling

Jan 2019 – May 2019 · 4 mos · Bengaluru, Karnataka, India

Jun 2017 – Dec 2018 · 1 yr 6 mos · Mumbai Metropolitan Region

• Built long only and market neutral quantitative trading strategies in R and Python
• Built end to end financial products which helped increase product diversity in the firm
• Work involved data extraction and cleaning, building frameworks for quick strategy testing, researching for new ideas and implementation
• Worked with Spark, Livy, Presto, Zeppelin and built trading frameworks using DataFrame in Spark with scala to run models on distributed data sets.
• Worked on various optimization techniques like Nelder Mead, Genetic Algorithm, Risk Parity etc.

May 2016 – Jul 2016 · 2 mos · Bengaluru, Karnataka, India

• Developed intuitive dashboard interface, which provides quick visibility to high risk SOP’s and vulnerable sites. It enables companies to capture, identify and report complex risks before they escalate to a catastrophic event. All the visualizations were done on Tableau.

May 2015 – Jul 2015 · 2 mos · Greater Lucknow Area

• GT-Silicon is a startup which has made an embedded device for pedestrian navigation system in collaboration with KTH Royal Institute of Technology, Sweden. I had worked on their device extensively during the internship to figure out solution to a technical problem present on their device. Magnetometer and pressure sensor present on their device were not functioning properly and I had to figure out an alternate way of programming registers. Work also involved calibrating data from various sensors and sending it to a mobile app.