How did you develop an interest in Machine Learning?

I had just graduated in Physics and was hesitant to embark on a long and scary PhD journey, when a friend asked me to help with some data analysis “because I could probably figure out the maths”. A few python and ML online courses later - I was hooked. Something about the experimental development process and the empirical nature of the methods felt exciting and powerful, and still does!

What have you been doing before joining Jungle Program?

My first love was Natural Language Processing, I was working with text analysis for legal AI. My favourite challenges related to the scale of these entreprise systems and how to reliably integrate performant models. Machine Learning was getting a lot of research hype - but what about the engineering side? In parallel, I was also a lecturer in Machine Learning at Ilia State University in Tbilisi. This turned out to be a bit ambitious on my timetable, but a delightful and rewarding experience. Most recently I co-founded Watergenics, where we are building the next generation water data platform.

How are you using advanced Machine Learning to create a water data platform?

We are using it for two main applications:
1 - Optical sensors produce spectrographic data that need to be translated into chemical quantities. ML algorithms can learn this mapping to improve the accuracy of the predicted concentrations. This is effectively powering-up water quality sensors with AI!
2 - Rivers and streams are environmental veins, and as such, the health of these ecosystems are reflected in the water quality. With a network of accurate water quality sensors, we can build predictive analytics to improve both the efficiency and sustainability of agriculture and aquaculture. For example, detecting diseases before they spread, or optimising crop yield.

What inspired you to pursue a career as a teacher?

During my studies, I became fond of scientific communication - how can one most succinctly convey complex ideas? Popular science authors are part data compression craftsmen and part telepathic magicians. A big chunk of it involves putting oneself in the audience’s shoes - and this is particularly true with students. I discovered this interpersonal aspect through tutoring, it felt like trying to reverse-engineer someone’s thinking! But beyond the brain puzzles, teaching is special because one gets to witness the results of the learning... There’s something immensely rewarding about seeing a student grow, and feeling like I contributed a little bit.

Why learn Data Science & Machine Learning today?

All industries are now inundated with massive amounts of data. The biggest challenge is turning this information into insights or features. Even my toaster can connect to wifi nowadays… and has data waiting to be explored! Data analysis is slowly becoming a universal skill rather than a niche proficiency. I’d like to see excel dethroned by pandas + sklearn as the status quo for quick data analysis. I’m also a general fan of empirical methods and statistics, and I fundamentally believe they are our best tools to understand the world. More Data Scientists and more Machine Learning Researchers means more people helping to interpret what’s happening around us.

What makes you a great teacher?

Mostly because I was a bad student. I used to skip university lectures a lot because they felt so anonymous and inefficient - we were 250 students in a large amphiteatre, and the pace of teaching had to fit with the questions and learning styles of too many. There were some lecturers I avidly followed however, and this made me appreciate what really clicks in a captivating and useful lesson. This has been reinforced in the age of the MOOC, since most educational content is now openly available online. The value and duty of the teacher must go beyond the simple sharing of content: it’s about effective guidance and learning guarantees.

Could you share with us your 3 favorite Machine Learning open-source projects?

One for the Data Scientists: streamlit. My favourite way to interactively communicate insights from exploratory data analyses.

Streamlit • The fastest way to build and share data apps
Streamlit is an open-source app framework for Machine Learning and Data Science teams. Create beautiful data apps in hours, not weeks. All in pure Python. All for free.

One for the ML Researchers: huggingface/transformers. Trend-setters when it comes to open access of pretrained models, and one of the biggest forces behind the NLP boom.

Hugging Face – The AI community building the future.
We’re on a journey to advance and democratize artificial intelligence through open source and open science.

One for the ML Engineers: ONNX. The kind of language-agnostic production thinking we need!

ONNX | Home

Click below to apply for the following bootcamps:

How to apply