Data Sanity Breakfast: Pet Projects Edition

This time, we’re putting pet projects in the spotlight!

Join us for a cozy morning meetup with fellow data professionals and enthusiasts. Share your project or just join to listen — we welcome both! Come for the coffee, stay for the stories ☕️

⚡️ 3 minutes to present a project + 2 minutes for Q&A.
⚡️ No project? No worries — come to listen, network, and enjoy coffee.

📍 Dusha Specialty Coffee, Belgrade
📅 Sunday, June 22
⏰ 10:00 – 12:00

✒️ To join, fill out the registration form

Data Sanity Breakfast: It’s Already Tomorrow!

Start your Sunday with a perfect blend of AI projects and coffee ☕️

📍 Dusha Specialty Coffee, Belgrade
📅 Sunday, June 22
⏰ 10:00 – 12:00

✒️ Sign up now via this short form

Thanks to everyone who came to Data Sanity Breakfast last Sunday! Apart from coffee, chat, and networking, we discussed several pet projects — here are just some of them ⚡️

stranger.news by @yarrrro
We could use LLMs for good — or to create a conspiracy theory website inspired by the Call of Cthulhu universe! Help wanted to incorporate user interactions and create a suspenseful story (web development and advanced AI workflows).
🔗 https://stranger.news/

Daily Insights by @an_klowait
Inspired by daily "cards" from a Mental Health Awareness marathon, this lightweight applet uses few-shot-prompted LLM calls to deliver quick 10-second mindfulness tips. Coded for fun over a weekend!
🔗 https://huggingface.co/spaces/an-klowait/daily-insights

Distilling Pretrained LLMs Into Recommender Systems by @nikis14
Distillation pipeline that uses pretrained LLMs to improve accuracy of a recommender system without a fine-tuning overhead.

P.S. We're already preparing several new events for you — stay tuned!

We're launching a workshop series focused on AI and Large Language Models (LLMs)!

Over the next few months, we’ll be running workshops every 2–3 weeks covering key topics, including:
🔹 Introduction to AI & LLMs
🔹 AI Agents
🔹 Retrieval-Augmented Generation (RAG)
🔹 Evaluation of LLM Applications

🎯 Next up: Introduction to AI & LLMs

Join us for theory, hands-on insights and practical tips.
If you’ve never worked with AI or LLMs before, don’t worry: this is the perfect and friendliest place to start.

✨ Register to secure your spot

🗓 Saturday, July 5 | 11:00 AM – 1:00 PM
📍 Belgrade, Startit Centar

Another recording from our June edition of Data Sanity Talks is out!

⚡️ How We Use AI Tools: Mental Models & Misconceptions by Ilya Zakharov

Learn practical frameworks for getting the most out of AI and avoid the traps that trip up most users.

Enjoy! ✨

⚡️ Data Sanity Talks, London Edition: Last Call for Speakers!

Speaker applications close this Sunday, 13 July.
If you have valuable insights for the DS/ML/AI community, we’d love to hear from you.

✏️ When: Aug, 6:00 — 9:00 PM
✏️ Where: London, Civo Tech Junction
✏️ Format: 15-40 minute talks + Q&A, professionally filmed and shared across our media channels

What we’re after:
- Technical in-depth explorations: AI/ML methods, models, tools, cutting-edge research insights
- Big-picture perspectives: project & product leadership, industry trends, ethics, strategy

Apply as a speaker & Learn more 👉 https://datasanity.dev/

Data Sanity Talks is coming to London — public registration is now open!

🗓 Date: 6 Aug, 6:00 – 9:00 PM
📍 Venue: Civo Tech Junction, London
🔗 Save your seat: https://datasanity.dev/

Why attend?
• Learn case studies from leading AI practitioners
• Connect with data scientists, ML engineers, and product leaders
• Enjoy complimentary snacks and coffee


⭐️ Join us on LinkedIn
⭐️ Follow us on Instagram

It's time to announce our first speaker for Data Sanity Talks, London edition!

🎙 LLMs Are Expensive. Here’s How Big Tech Makes Them Less So

Speaker: Max Shaposhnikov, Research Engineer @ Tessl, ex-Applied Scientist @ Amazon Research

In this talk, we'll explore
- How long and how much it really costs to train an LLM – back of the envelope calculation framework
- Techniques to extend context windows with increase in quality
- Modern Speculative decoding strategies
- Fast, production-ready guard models, like those evaluated in the recent CircleGuard benchmark, that make safety checks viable at scale.

🗓 When: Aug 6, 6:00 — 9:00 PM
📍 Where: London, Civo Tech Junction
🚀 Register & Learn more: https://datasanity.dev/

It's time to announce our next speaker!

🎙 LLM Super‑Teams: Modern Ensembling Strategies That Beat Any Single Giant by Polina Proskura, Applied Scientist @ Amazon


Here’s what we’ll uncover with Polina:
- The challenge of uncertainty in real‑world LLM applications
- Ensembling strategies to boost robustness and reliability
- Methods to quantify and reduce model uncertainty
- Cost‑effective approaches for deploying ensembles in production


🗓 When: Aug 6, 6:00 — 9:00 PM
📍 Where: London, Civo Tech Junction
🚀 Register & Learn more: https://datasanity.dev/

🌟️️️️️️ Fresh recording from our June edition of Data Sanity Talks!

🎙 Hacking the Prompt: AI Security in the Era of LLMs by Miloš Božić, CEO & Founder at Pyxero.ai

In this talk, Miloš shares how techniques like prompt injection, system impersonation, and chain‑of‑thought hijacking can bypass safety filters and exploit model behavior. You’ll learn where current defenses fall short, and how to make AI more trustworthy.

⭐️ Join us on LinkedIn
⭐️ Follow us on Instagram
⭐️ Visit our site

Let’s unveil our third speaker for Data Sanity Talks, London Edition!

🎙 Navigating the Enforcement of the EU AI Act: Practical Insights for Compliance by Michael Borrelli, Director @ AI & Partners

In his talk, Michael will explore:
- the enforcement mechanisms of the EU AI Act
- the importance of audits and risk assessments, and the implications of non‑compliance
- concrete measures for proactive compliance
- ways to build robust governance structures before enforcement takes effect


🗓 When: Aug 6, 6:00 — 9:00 PM
📍 Where: London, Civo Tech Junction
🚀 Register & Learn more: https://datasanity.dev/

Data Sanity Talks: London Edition is almost there — here's our speaker number four!

🎙 Accelerated Diffusion Models via Speculative Sampling by Alexandr Galashov, Senior Research Scientist @ DeepMind

Here's what Alexandr will show us:
- How to maintain generation quality with fewer steps
- Bottlenecks in diffusion‐model sampling
- How full posteriors help with faster sampling

🗓 When: Aug 6, 6:00 — 9:00 PM
📍 Where: London, Civo Tech Junction
🚀 Register & Learn more: https://datasanity.dev/

Call for Speakers: Belgrade, October 4!

Data Sanity Talks is coming to Startit Centar on Saturday, October 4 — a full day for data scientists, AI practitioners, and industry leaders to connect, learn, and share ideas.

Apply to give a 15-40 minute talk with slides on AI, Data Science, or Data Engineering. Possible topics:

🚀 Deep technical dives: ML/AI model architectures, tooling, the latest research breakthroughs, and best practices in ML system design.

🚀 Broader overviews: Leadership and management strategies, industry trends, and ethical considerations.

🔗 Apply to become a speaker
🗓 Deadline: August 17

It's time to announce our final speaker for Data Sanity Talks, London Edition!

🎙 Data Sorted! How to Turn Mayhem into Insight by Oliver King-Smith, Founder and CEO @ smartR AI

Takeaways you won’t want to miss:
- What assistive intelligence is
- How to use it when transforming messy data into actionable bits
- A case study to show an example


🗓 When: Aug 6, 6:00 — 9:00 PM
📍 Where: London, Civo Tech Junction
🚀 Register & Learn more: https://datasanity.dev/

Data Sanity Talks, London Edition is happening tomorrow! 🎉
We’re excited to see you. Join us for an inspiring day!

🗓 Date: Wednesday, 6 Aug | 6 – 9 PM
📍 Venue: Civo Tech Junction, London
🔗 Reserve your spot: https://datasanity.dev/