My first experience as a mentor during “Artificial Intelligence Labs”

AiLabs was my first experience as a mentor. Usually I’m a participant but I had this time the chance to be on the other side, representing Deepomatic.
This event was kind of like a startup weekend but the use of AI (Artificial Intelligence) was meaningful, and most of the participants were already working in companies.

Thursday: Bootcamp

Thursday evening was a bootcamp about AI-related business and the tools some partners of the event were providing. The first speaker was Louis Dorad (General chairman at Papis.io), he explained how to use your data to create an AI (ML) project using the “machine learning canvas”, a method loosely inspired by the world famous “business canvas”. The second speaker was Yacine Rezgui (IBM evangelist), he showed us a pretty cool demo about an IBM Watson tool that gives computers the possibility to read text with emotions. The second part of the demo was how to use Node-Red, a kind of drag & drop tool to create a conversational pipeline with natural language processing (NLP) and IBM Watson. Unfortunately the demo failed (that’s why you always need to have a video of your demo just in case!). The third speaker was Hicham Tahiri (CEO of smartly.ai), he first explained what smartly.ai is and then what was going on with chatbots. Last but not least, it was Augustin Marty’s turn (CEO of Deepomatic). You can read this great blog post by Vincent Delaitre (CTO of Deepomatic) explaining what you can do using our APIs. Furthermore, for this event we had already some databases ready to use in various fields (posters, furniture, books…). The bootcamp ended with a friendly “wine & cheese” (French of course).

Friday: Team creation & first pitch

During the entire weekend we were at Blablacar’s offices: the place is amazing, nicely decorated and well located in Paris. I reached the location on time for the 24 opening pitches, some of them were realistic, others were fanciful (trying to overtake Google) and others were too mysterious for me to have a good opinion on them. One pitch was: “I want to revolutionize medical research using AI”. Interesting, but how? One of the pitches that caught my attention was by Looked because they clearly needed to use our APIs. After this session everyone could vote for three teams in order to keep 10 teams. Unfortunately Looked was not one of the ten. After some pizzas each team got a room for them to start working. It was a bit too early to visit every team to see if they needed help. However, we spoke with the team Tipi, whose original idea was to create an app for people looking for apartments. The interesting point was that they were willing to use our similarity API to “simulate” regression and get information about the quality of an apartment based on the images. Globally (from my point of view), at this point all the teams were focused on the business part, browsing Techcrunch and similar websites to find a sustainable business model.

Saturday: The adventure continues

Saturday I went to the event with Vincent. My first impression was that all the teams had worked a lot and had clearer goals. One of the teams that we exchanged with was the team “Her”, their purpose was to create an AI that discusses with lonely people in order to detect emotions. The pitch was nice but as Vincent said:

Keep in mind that you will have to convince a jury of VCs. Having a vision is important but the path to this vision will be long and full of unexpected (good and bad) surprises. So this weekend is more about defining your first minimal product, the people you target and a way to reach them.

We talked with other teams, mentors and organizers. After a while, I went to visit Tipi to see how the project had evolved since the prior night. The main target for the project were the people willing to buy a flat and no longer the people looking for a place to rent. I was happy to help the team to crawl some data in order to index them in our database and perform a similarity request. It was nice to see people from the “outside” using our product that was exclusively internal before, especially since their feedback was nice: “Easy to use, producing good results, …” We also saw the first mock-up.

Sunday: Final stretch

Sunday was the day for the final adjustments. All the teams were prepared, some of them asking last minute questions, some of them practicing their pitch in front of us. The morning was booked for a session of “white pitches” in front of the mentors to get some feedback, and then the final pitch session was at 4pm. The winner was the team Enhanced politics, the runner-up was Tipi.

Conclusion:

Personally, it was an awesome experiment, I met and discussed with many interesting people. I explained and saw a group using our APIs for an all-new usage and it was working very well! It was also nice to be able to spend more time on the vision and the business part of all these projects (usually I just participate in the hackhathon where it’s mainly coding and doing something fun).

Of course a big thanks goes to the team of AIlabs, for the idea of this event, gathering all these people in nice places (42 and Blablacar), getting a glimpse of the behind-the-scenes. I really believe that we are starting a new technological revolution and it’s marvellous that people work on this kind of event in order to create an even better tech-friendly environment in France.

Alexis Jacob, Backend Engineer @deepomatic