About Me

Keeping up with recent trends, I have been learning about AI Agents with Hugging Face's latest course, while trying to stay grounded and rediscovering Ruby on Rails and beauty of "choosing boring technology". In my free time, I have been dabbling with game development in Godot, and published a couple of simple minigames.

Outside of tech, I recently obtained the Level 2 WSET Award in Wines and am now preparing for Level 3. I have been learning the ins and outs of making gelato (@gelato_sottovoce) and obtained a Food Sanitation Supervisor license, hoping to run a pop up store someday.

I am still an avid language learner: I have obtained JLPT N1 in 2021 but still practice my kanji daily.

My return to Indeed was unexpected but rewarding: I wanted to keep building products with AI/ML, and when a role came up for a PM on the Recommendation Science team, with the mission of creating a matching marketplace for jobs, it sounded like the perfect fit.

As a very technical team powering a variety of products, my main contribution was to translate business goals into a systematic roadmap for ML model improvements. We helped our client teams scale to 20 markets and double conversion rates to become one of the best channels for users to find jobs.

I learned that ML teams have different cadences and needs from traditional product development, and adjusted our frameworks and processes to prioritize product impact while still investing in continuous innovation. This hybrid charter allowed us to pioneer and scale new technologies at the company, including Deep Neural Networks and Multi-Target Models, and grow into a 20-people team spanning R&D and ML Ops.

Highlight: the pride of seeing my team publish 5 papers at RecSys and RecSys in HR 2024.

While learning the ropes as an engineer, I focused on "how" to build products, but I also wanted to participate in defining the "why" and "what", and experience building products across industries.

As an "expert consultant", I got to blueprint and prototype new services in healthcare and finance, boost conversions on e-commerce sites, and build analytics solutions to help engineers optimize site yields using Machine Learning, creating value at each iteration thanks to the Agile methodology.

I then transitioned to a generalist role, and dove deeper into digital strategy, where I helped tech players select the best opportunities to grow in Japan, and conducted tech due diligences to help inform startup M&As on software quality and team maturity.

Highlight: being a co-author on Japan Digital Agenda 2030, a paper from McKinsey and the American Chamber of Commerce in Japan which lays out a path for digital transformation for the Japanese economy and society.

Moving to Tokyo was an unexpected joy. I had begun studying Japanese in the final year of university, and jumped at the opportunity to live in Japan when I came across an opening for new grads at Indeed. Back then, I did not expect this adventure to last this long.

At Indeed, I worked on integrating and developing job search sites in our portfolio, using them as a test bed for SEO optimization, as well as product and tech stack innovation: we were among the first teams to adopt Kotlin and ReactJS, alongside Java and JavaScript with jQuery.
During my time, I sharpened my skills in A/B testing, implemented an identity, profile and resume management system as one of our first AWS use cases, and built a frontend test suite in Python to automate the verification of our portfolio sites.

Highlight: spending a week-long spike prototyping the email-based recommendation product which eventually grew into "Invite to Apply", of which I became the PM in 2021.

I began my university journey as a Linguistics undergraduate, excited to tackle the domain of human language with a diverse toolkit of humanities and sciences. In my second year, I came across Natural Language Processing, and was fascinated by the intellectual breadth and potential of Computer Science.

I decided to make the jump, and spent the summer as a research intern at Linguamatics while covering the first-year CS coursework to make that happen.

As a "CompSci" I extended my interests from NLP to general AI, as well as Information Retrieval, Distributed Systems, and Complexity Theory. Bridging my two academic paths, I wrote my final dissertation on efficient language parsers.

Highlight: winning the "Hack the Brain" hackathon with WinkIt, a mobile app that uses brain signals to control your smartphone.