Jan 2025 – Nov 2025 · 10 mos · Singapore · Hybrid

• Neoflo emerged from hands-on experimentation with generative AI, where we identified early potential in applying the tech to repetitive, human-reliant operational work.
• As co-founder and CTO, I was responsible for research & development, including end-to-end systems design, architecture, workflow orchestration, model selection, and operational tooling. In a nutshell, cross-pollinating two decades of distributed systems knowledge to agentic systems.
• The primary learnings from early pilots revolved around the practical challenges of unstructured document processing, including multi-language scenarios (with non-Latin character sets), tradeoffs between OCR versus text- and vision-based models, and the efficacy of small models (over large models) for narrow tasks. Solving for reliable and efficient context-window management during workflow handoffs, tool-calling, and human-escalations proved to be worthwhile investments. Additionally, production-grade observability, evaluation, and failure-detection and -handling to arrest hallucinations were paramount for real-world applications.

Sep 2019 – Dec 2024 · 5 yrs 3 mos · Greater Jakarta Area, Indonesia · Hybrid

• I co-founded Ula to address structural inefficiencies in Indonesia’s traditional retail trade. Small retailers in Indonesia were plagued with a fragmented supply chain leading to inconsistent product availability, and fluctuating product pricing coupled with limited access to credit. Coming from an e-commerce background from the West, these were fresh challenges for me.
• At peak, Ula hit a $300M sales run rate, processed 500,000 orders per month, and served over 100,000 mom-and-pop stores across Indonesia, with best-in-class product selection and pricing, warehousing, logistics, and embedded credit.
• As CTO, I was responsible for technology strategy, systems architecture, and execution across Product, Engineering, Design, and Data, directly managing a team of ~200 across Singapore, Indonesia, and India. I also supported multiple funding rounds, contributing to raising over $140M from global institutional investors.
• From a systems perspective, I developed architectural and operational practices tailored for Indonesia’s real-world constraints, including unreliable networks, low-end mobile devices, and unpredictable last-mile conditions. This included building internal logistics routing and decisioning systems where off-the-shelf tools proved insufficient. Separately, I drove initiatives to aggressively adopt AWS spot instances to reduce cloud infrastructure costs and to increase system robustness by designing resilient, idempotent systems that expect instance-loss as a normal operating condition.

May 2018 – Jul 2019 · 1 yr 2 mos

• I was responsible for leading the technology function in the BookingSuite department. Practically, this meant ~70 people in cross-functional teams: developers, designers, data scientists, analysts, copywriters, and quality specialists). My role was focused on mid-to-long term initiatives, for example quarterly tech roadmaps to ensure continued runway for innovation, to co-creating the 3-5 year strategy for the department with the Product organization. I was also responsible for the growth and resiliency of the department through hiring, employee engagement, talent development, organizational design, and developing engineering culture through leading-by-example.

Oct 2017 – May 2018 · 7 mos

• I was managing teams across three distinct Hotel Tech products (~35 people in multi-disciplinary roles). The products ranged from a Hotel Guest Management System, to an insights tool for setting nightly rates for hotels, to a direct-channel website builder geared to the Hospitality industry. On a technical contribution perspective, I provided guidance to development teams to evolve systems architecture, balancing scalability, while minimizing technical debt, and maximizing velocity. My role allowed me to work closely with the Department Leadership Team and Product organization to align on OKRs/KPIs, to execute on a choreographed strategy.

Apr 2017 – Oct 2017 · 6 mos

• I managed three teams that owned the end-to-end software stack that allowed a Hotelier to create a property website to enable direct bookings (this included the Partner website creation experience; the Guest browsing experience; and the Guest booking experience). I was responsible for coordinating teams in Amsterdam and Seattle to move the entire AWS-based software stack into Booking’s existing infrastructure and data-centers, while maintaining availability of our Partner’s websites.

Jan 2017 – Apr 2017 · 3 mos

• Managed a team of six that owned the Guest experience of a Hotelier’s public-facing website.

Nov 2015 – Dec 2016 · 1 yr 1 mo · Greater Seattle Area

• Brought into the company as the sole Software Engineer to focus on the design and implementation of the core data pipeline for operational data in the field. WISErg employs network-connected machinery in the field (think industrial-grade IoT devices) to process food scraps. The data pipeline allows for real-time aggregation of this vast array of operational data and presents it for business consumption. Various applications were built on top of this platform to provide operational insight, operational history, as well as operational optimizations such as route-planning for the service team.

Jun 2014 – Oct 2015 · 1 yr 4 mos · Greater Seattle Area

• Hired as the second Engineer for Atlas, a product to "remember everything seen on any device and app while providing personalized recommendations". My focus was on building the backend infrastructure (primarily, in Node.js): from efficient inter-service data routing and RPC, to a first-cut user authentication service backed by Amazon Dynamo DB, to a prototype persistence store backed by Mongo DB and Amazon S3. As with all start-up opportunities, my reach extended beyond the backend systems, with forays into frontend installers using NSIS, and early user-facing prototypes using WinForms and WPF.

Jan 2013 – May 2014 · 1 yr 4 mos

• Served as the technical lead for rearchitecting a core Tier-1 system to ensure long-term scalability to support explosive business growth. The core of the project is a critical, high throughput (1B+ requests/day), low latency (<10ms) distributed system.

Mar 2009 – Jan 2013 · 3 yrs 10 mos

• Actively contributed to an event-based pre-computation framework used by over 200 teams. Later, by migrating the team's software to run within this framework, the team saw a $1M+ reduction in hardware costs.

Jul 2006 – Mar 2009 · 2 yrs 8 mos

• Reduced shipping costs, by implementing an inventory transfer system where inventory would be predictively forwarded to fulfillment centers closest to the areas in which demand was expected. Along similar lines, implemented portions of a another system to reduce inventory held by normalizing inventory across the fulfillment network.