Bodygram provides businesses with unparalleled opportunities for personalization by leveraging the body measurements of their customers.▶ Drove a cross-functional 20-members team to design and deliver the sizing solution adopted by apparel giants like Uniqlo and GU, reaching 10M+ MAU and increasing Bodygram's ARR from 150K to 1M in a year.▶ Led the development of Bodygram Platform, providing businesses with patented technology to obtain precise (<5mm MAE) body measurements from two RGB photos of their users.▶ Engineered and delivered highly secure GDPR compliant stacks, attaining the globally recognized and industry-leading ISO 27001 certification.▶ Overhauled the engineering department, establishing clear principles, and drastically boosting deployment rate from 2 releases per month to 10+ per day.▶ Led Bodygram's recruitment effort, effectively doubling the engineering team size from 10 to 20, and drawing talent from renowned companies such as Amazon, Google, and Stripe.

As Tech Lead for Google's Cycling team, I was responsible for identifying and designing solutions to improve the quality of cycling directions worldwide. ▶ Spearheaded the development of Naturalistic Routing, leading to substantial improvements of driving directions and to one of the world's largest applications of reinforcement learning with 100M+ states.▶ Designed complex anonymization pipelines processing 100TB+/day of highly sensitive location data to produce training sets and key performance metrics (see below). ▶ Aligned multiple teams (spanning 40~50 members) towards common objectives by designing and establishing responsive quality metrics to assess the quality of cycling directions. ▶ Secured buy-in from Google Maps leadership to further invest in Naturalistic Routing, leading to the creation of a specialized team focused on extending Naturalistic Routing to other transportation modes.▶ Tackled significant innovation roadblocks by partnering with the privacy and legal teams to establish and adopt new anonymization methodologies within the Google Maps organization.

As one of the four founding engineers of the Google Travel Transit team, I played a pivotal role in creating a new vertical in the travel space. Operating with a startup ethos within Google, our team rapidly developed and launched a service enabling Google Pay users (and later Google Search users) to search and book train tickets in India. ▶ Successfully established a new vertical within Google Travel in less than a year, leading to the most popular micro-app within Google Pay in India and helping 100M+ users to book the best trains.▶ Architected and delivered a highly-robust booking and search engine, processing 300K+ ticket purchases per month while providing low-latency (<50ms) access to ticket fares and availability globally. ▶ Saved Google $70M per year in third-party API fees by identifying the need, and delivering an ML powered distributed cache for ticket availability.▶ Contributed to 10+ tactical features, achieving a cumulated 16% increase in booking rate, positioning Google as the second online train travel agency in India.▶ Mentored 4 junior engineers who progressed to senior roles. ▶ Routinely partnered with a large group of stakeholders and leaders (Eng, PM, UX, BD, and Legal) globally distributed (Switzerland, India, Singapore, Australia, and US).

As a member and later Tech Lead of the Travel Logs team, I led the design and development of data pipelines processing hundreds of TB daily, extracting vital insights for Google Flights and Hotels. Recognized as the "go-to" person for Travel logs analysis, I was instrumental in aiding cross-functional teams to understand and enhance product performance.My key achievement was the creation of 'Travel Sessions', a comprehensive data warehouse for the Travel vertical. Travel Sessions streamlined multiple data pipelines, reducing processing costs significantly, and enabled Product Managers and UX Researchers to perform self-service analysis. This shift not only saved costs but also freed Engineering resources from performing routine analyses.

As a 20% contributor to Google Research team, I focused on integrating cutting-edge scheduling algorithms from my PhD research into Google OR-Tools. This integration not only strengthened OR-Tools' capabilities but also reinforced its position as a leading solution to tackle the world toughest problems in combinatorial optimization. Indeed, Google OR-Tools had been winning gold medals of international competitions since 2013, and my contributions played a significant role in maintaining and extending this winning streak.Specifically, I focused on lazily encoding the deductions of scheduling propagators into clauses and predicates. These are then leveraged by the SAT solver (at the core of OR-Tools' Constraint Programming solver) to strengthen its inference and learning capabilities thus significantly reducing the size of the search space to explore. Some of my other contributions include designing a property-based testing framework for various advanced filtering algorithms, and designing/implementing a new task-specific sorting algorithm that achieved ~30% speed improvement compared to the default option.

As a FRIA research scientist, I conducted fundamental research in computer science and combinatorial optimization, designing algorithms to solve NP-complete problems that were previously too large to be solved in practical time. Many of these algorithms were implemented in commercial/professional optimization tools such as Google OR-Tools or IBM CPLEX.Beyond my research work, I taught advanced computer science to classes of around 50 master's students, ensuring a comprehensive and engaging learning experience. I also played a key role in supervising Master's theses, where I guided students through complex research methodologies and helped them deepen their understanding and mastery of intricate concepts in the field.