My mission is to support Feetobook in adding ML-powered features to their products while developing AI expertise internally.
I draw from my various experiences as a team leader, an assistant professor, and a start-up founder to strategically decide where and how to inject AI culture and skills into an experienced SaaS company. The challenge is to define suitable projects that both make business sense for the company and are accessible to a team with no prior AI experience or expertise.
We started with a 12-week programme on a first project with a single engineer that allowed me to understand the current product and data available, which led to a first feature deployed into the product.
Following this initial project, I performed an AI audit to identify potential follow-up projects and their level of complexity. After many discussions with the founders, the company dedicated a team of four engineers to work on AI-related projects under my lead.
We defined one main project, which required retraining the team in Reinforcement Learning, for which we relied on online resources supported by weekly sessions grounding our learnings into the project at hand. After 2 months, the team had sufficient background to tackle the project via quick iteration, working progressively from simplified simulated versions of the problem up to the use of production data within a user-facing product. The final product is built from the ground up by the team, is tailored to their business and data architecture, and makes financial decisions directly impacting the bottom line.
In parallel, we worked on smaller projects using LLMs and Computer Vision tools that contribute to establishing an AI culture within the company as they are developed into products and progressively involve other business units in design, UX/UI, and marketing.
Interestingly, all this is achieved while only spending half a day a week at the company. We make progress on each project via a mix of all-hands and focused 1:1 meetings, with clearly defined weekly goals.
