#7 Berlin AI Community Update

Ha Duong
Applied Artificial Intelligence
4 min readFeb 6, 2018


First of all, we want to introduce our extended Berlin AI team connecting AI practitioners, sharing expertise and challenges on #appliedAI:

Minh Ha Duong, Investment Manager at Project A who’s generally excited about nerdy deep tech stuff;

Tamara Bobic, AI & NLP Engineer at Parlamind who’s a years long researcher and developer on NLP, machine learning and semantic technologies;

Appu Shaij, Head of R&D at EyeEm who people call a superhero in computer vision and programming

Ha, Tamara, and Appu

Insightful read! — What grabbed our attention the past weeks has been the insights from Susanne Westphal, data scientist at Berlin-based machine vision startup TwentyBN on using synthetic data for deep learning video recognition

Congrats! — Furthermore, it’s exciting to see the automated self-owned forest project Terra0 at the Berlin University of Arts being mentioned by LA-based Colab partner Arthur Kiulian in a recent Forbes article

Berlin is on the global map! — Also ‘jobbot’, the next-gen job discovery chatbot MoBerries, which has been named as 1 of the 100 hottest startups in Europe by WIRED, is one of the 10 most used messenger bots worldwide, has become freely available on Facebook’s Messenger

Exciting for Berlin as AI hub! — End of 2017, Deckard.ai which brings machine learning and process mining to project management such as automating task effort estimation, completion-time forecasting, resources allocation and roadmap planning, has been named 1 out of 5 hottest startups in Europe through Nvidia’s Inception program.

Keep going !— We also recognized that that the AI-startup understand.ai closed its funding round, the expansion of Brent Hoberman’s co-founded ‘Entrepreneur First’ to Berlin program where education on how to build successful companies with future technologies such as AI are key, and the public launch of Ocean Protocol’s ‘Decentralized Data Exchange Protocol

Shared Lessons Learned @ Berlin AI #7

Last week, Berlin had it’s #7 applied AI community meeting hosted at Spielfeld Digital Hub (Thanks for hosting us!) with AI practitioners from early-stage startups such as Candis and Terraloupe, growth stage ones such as Finiata, and large tech companies such as Zalando. After starting almost 2 years ago, Berlin has shown to not only have a vibrant, but really engaging AI community that is actively sharing lessons learned to enable others apply AI from leading AI practitioners such as Ralf Herbrich, director of machine learning at Amazon, Reiner Kraft, VP of engineering for search & personalization at Zalando, Martin Karlsson, CEO at Lateral, Rasmus Rothe, CTO at Merantix, and many more.

This time, Martin Schmitz, head of data science services at RapidMiner.com, a data science software platform, shared his lessons learned on selling data science and why you should ignore data science performance measures for that. The most interesting lessons learned was that data science performance measures cannot be used by themselves to justify business decisions. When selling to internal or external customers, make sure to communicate the potential business impact (derived from your data insights) in order to incentivize decision-making.

See our video recording here (apologies for issues, technology let us down this time)

Followed by Baver Acu, head of technology at GermanAutolabs.com, the digital assistant for automotive, shared insights on their challenges and learnings from building an AI co-driver. What was interesting for us to learn was that adding AI capabilities (voice, visual action recognition) to a small hardware device needs to respect the constraints in terms of processing power and battery life time.

Martin and Baver

Thanks to those that joined us and we hope to see you engaging in future too!

Presented AI Clinic challenges

Please get in touch with berlin@city.ai if you can help the below AI practitioners further solving their individual application challenge

From Smart Cities to Aware Environments, by Massimiliano Bariola

Can a physical space from a corridor to a city be made aware of what happens within it? can it intelligently interact with other such spaces as organs of a more complex organism, an AI ecosystem, coordinating for a higher level goal?

If you have one yourself, submit it here to present yours at our next Berlin AI community meeting or ask for relevant introductions to peers globally.

Other AI activities in Berlin

If you want to share or co-host your AI related event or activity with us, we’re eager to support you. Simply email us at berlin@city.ai

Next Berlin AI Community Meeting around April 2018 — If you haven’t sign up yet, do it here so we can notify you 3 weeks in advance.

Let’s enable everyone to apply AI, your Berlin AI team

Appu, Tamara, Ha, Manuel, Ramzi, Christoph



Ha Duong
Applied Artificial Intelligence

Investment Principal at Ocean Investment (single family office). Advisor at Cambrial Capital, Mentor at Techstars, Contributor at Forbes. http://minhhaduong.com