On Tuesday, February 25th, we’re hosting a new edition of our 100+ free AI events in 2026 series, starting at 6 PM at Startit AI Hub in Belgrade.
Main topic of the evening: what we can do to make our AI agents reliable and their responses accurate. Bonus: how context engineering works in practice, using tools like Claude Code and Copilot CLI as examples.
— Engineers who want to build AI agents and tools
— Teams building LLM-based systems
— Anyone using Claude Code, Copilot CLI and similar applications who’s curious about how they work
Our guest is another familiar face, Aleksandar Anđelković, Software Engineering Manager at Microsoft. You may have seen him at our AI Week event last year or in the podcast we recorded this summer.
Anđelko works on Context IQ — a component that enables Copilot to understand context in Word, Excel, Outlook, and Teams. Before Microsoft, he led a third of the React team at Meta.
In previous events, we talked with Anđelko about grounding in general. This time we’re going deeper:
— Why increasing the context window doesn’t solve the problem
— Four steps of context engineering: Write → Select → Compress → Isolate
— What’s changing: from needle in a haystack to combining information from multiple sources
— Copilot CLI vs Claude Code: deep dive into the first, analysis of the second, as examples for the topic of the day
Fun fact: while at Meta, he met Boris Cherny, the author of Claude Code.
As always, we’re leaving plenty of time for questions and discussion with the audience.
Format: ~60 minutes of conversation + Q&A
Language: English
Attendance is free, but registration is required:
On Wednesday, February 18th, we’re hosting another event in our 100+ free AI events in 2026 series, starting at 6 PM at Startit’s AI Hub in Belgrade.
Main topic of the evening: what do open-weights models offer us these days, what actually happens under the hood when you run an AI model — and what it’s like to build open hardware for them to run on.
— Anyone interested in a deeper understanding of how LLMs work
— ML, backend, DevOps, SRE engineers working with AI infrastructure
— Open-source fanatics
Our guest is an old friend, Nikola Cvetković, kernel engineer at TensTorrent.
Nikola is part of the team that writes low-level code enabling AI models to run on Tenstorrent hardware — from pioneering models like the early Llama to newer architectures like DeepSeek’s.
At previous events we talked with Nikola about this topic in general and what TensTorrent does, and now we’ll cover a deeper agenda:
— the journey from complete dependence on OpenAI and others to today’s open-weights revolution
— going a bit deeper than usual — what happens at the hardware level when we run these models
— why this makes building open hardware much harder than building models
— why and how TensTorrent does it with RISC-V
As always, we’re leaving plenty of time for questions and discussion with the audience.
Format: ~60 minutes of conversation + Q&A
Language: English
Attendance is free, registration required:
On Wednesday, February 11th, we’re hosting a new edition of our 100+ free AI events in 2026 series, starting at 18:00 at Startit’s AI Hub in Belgrade.
The main topic of the evening is what it’s like to work on a search engine inside one of the fastest-growing AI startups, Perplexity — from the perspective of an engineer who does it every day.
The evening’s guest is Nikola Trifunović, Search Engineer at Perplexity. Nikola is a big tech veteran who spent nearly 15 years at Microsoft, and now he’s building the infrastructure behind one of the most popular AI products in the world.
We’ll discuss the following topics:
As always, we’re leaving plenty of time for questions and discussion with the audience.
Format: ~60 minutes of conversation and Q&A + networking, the event will be held in English.
Participation is free, with mandatory registration:
On Wednesday, February 4th, we are organizing a new meetup starting at 6 PM at Startit’s Belgrade AI Hub. This event is part of a series of 100+ AI Hub events in 2026 that are helping the IT community successfully adapt to the AI transformation.
The guest of the evening is Nikola Kolevski — Head of Engineering at the domestic fintech company Solflare, and the topic is: what makes engineers good in the AI era, how to recognize it in an interview, with a discussion on the application of AI in a mature yet still ambitious technology organization.
As Nikola tells us, reasoning models became serious in mid-2025. Since then, technology teams can no longer ignore the question of whether we are using AI, and consequently, how this is reflected in the selection process.
He and his team went through this process firsthand. Nikola will share concrete experiences: what worked, what didn’t, what lessons they learned, and what we still know and don’t know when it comes to AI in production, both literally and metaphorically.
Tentative agenda:
Format: ~60 minutes of discussion + Q&A
Language: English
On Wednesday, January 28, we’re hosting a new edition from the series 100+ free AI events in 2026, starting at 18:00 at Startit’s AI Hub in Novi Sad.
The main topic of the evening is how DevOps is changing AI today, how it will do so in the coming years, and we’ll also touch on how to build an AI-first startup.
Our guest is engineer and entrepreneur Darko Fabijan, founder of the new startup SuperPlane. Darko has deep experience in this space as the founder of Semaphore, a DevOps company that has been operating for 15 years.
We’ll discuss the following topics:
As always, we’ll leave a large portion of the time for questions and discussion with the audience.
Format: ~60 minutes of conversation and Q&A + 60 minutes of networking Language: English
Participation is free, but registration is required. We’ll publish the registration link soon.
On Wednesday, January 21, we’re hosting a new AI Hub meetup starting at 6pm at Startit’s AI Hub in Novi Sad.
Our guest is Stefan Trkulja — a designer from the “full-stack” school with ~25 years of experience. The topic: the present and future of the design role in the context of AI-driven transformation.
Over the last three years, design work has fundamentally changed: the tools we use, the speed of decision-making, and the expectations teams and organizations place on design roles — all of it shifted at once.
We’ll unpack that with Stefan, grounded in his experience working at an AI startup based in San Francisco — the place where many of these role shifts happen first, and then eventually reach the rest of us.
Rough agenda:
Format: ~60 minutes conversation + Q&A.
This event is for designers who work (or want to work) in the IT industry — and for anyone who wants to collaborate with designers more effectively or hire them.
Participation is [FREE / REGISTRATION REQUIRED].
On Wednesday, January 14th, we’re hosting the first AI Hub event of 2026, starting at 18:00 at Startit’s Belgrade AI Hub.
The topic is one of the most in‑demand roles in tech right now — Agentic engineer.
With Dragan Spiridonov, the Agentics Foundation representative in Serbia, we’ll discuss what this role entails and the discipline in general.
Who is this for
- Architects,
- full‑stack engineers,
- technical QA engineers,
and other IT professionals (frontend, product, devops) who want to upskill in line with new market expectations in the AI era.
The goal is for you to leave better prepared to use AI tools and a new way of thinking, so you can build more competitive products or become more valuable to your employer.
The problem we’re tackling is simple: old software and old development processes are becoming uncompetitive.
We already feel it, and soon we expect even more of the market’s money to move toward AI software and professionals who know how to build it.
No matter which recruiter you speak to or which research you read, you’ll see that new and still loosely defined roles like AI engineer are among the most in‑demand in 2026.
With Dragan Spiridonov we’ll discuss the Agentic Engineer role as defined by the Agentics Foundation — what it is, why it’s useful, and how
it can be learned.
Dragan has 30 years of experience in IT, has organized and spoken at QA conferences, and we’ll roughly follow this agenda:
As with our other events, we’ll leave plenty of time for audience Q&A.
Thanks to Solflare, a top fintech team in Serbia, this and other AI Hub events are free to the public.
Just register via the form below and we’ll see you on Wednesday.
