ML Debugging Workshop
A live 3-hour, hands-on session with me on how to effectively debug machine learning systems. Give your team the mindset, tools, and framework to accelerate project delivery and reduce uncertainty from unexpected bugs in your pipelines.
Why care about ML debugging?
Late ML initiatives. Slow progress. Rising costs, but little to show for it. When you ask what's blocking progress, the answer is vague: “The model isn't working, and we don't know why.”
In teams that use ML continuously, small issues compound. Experiments become harder to interpret and iteration slows down. Time-to-deliver not only grows, but also becomes more uncertain. Modern ML systems have many interconnected failure points. Without a systematic way to debug them, teams fall back on trial and error.
This workshop gives your engineers a structured, repeatable approach to diagnosing ML failures and learning from them. Higher quality and quicker, more reliable deliveries.
Your team will:
- Prepare effectively before issues appear.
Smart preparation has a multiplicative effect on debugging effectiveness. We'll cover practical preparation steps that set teams up for success: starting simple and expanding in controlled steps; following good coding practices; being closely familiar with their data; always having solid baselines to compare against, and more.
- Apply a structured, diagnostic approach to debugging ML systems.
Most ML engineers fall victim to availability bias, and end up chasing the first idea that comes to mind, usually spending more time than necessary. Instead, we'll cover how to systematically narrow down the cause, starting with quick, low-cost tests to rule out broad classes of issues, then progressively zoom in with deeper, more targeted investigation.
- Transform the bugs they get into learning experiences.
Learn how to reflect and adapt the framework to your own projects, so each participant leaves with a tailored strategy that fits their team's data, infrastructure, and workflow.
- Act on the knowledge they receive.
The workshop contains two thematic case studies where they'll immediately put what they learn into practice.
Past participants
“I really appreciated the workshop! What Juliano describes feels so obvious once he puts words to it, yet I haven't consciously thought about it before. He manages to concretise the hidden processes we all rely on in this field, and by articulating them, I feel more aware and intentional in how I debug and reason about ML problems.”
Anna Nylander, Smartr
“Great workshop that gave me new tools to debug machine learning models. The most interesting parts were how to prepare to minimize debugging time and the importance of reflection after solving a bug. Two things I take with me into my daily work.”
Robert Nyquist, Smartr
“Juliano identifies pain points we all encounter in ML work and provided practical strategies to avoid them. What resonated most with me was the mental shift: treating debugging not as firefighting, but as strategic information gathering. A game of maximizing insights while minimizing wasted effort.”
Fredrik Ring, Smartr
“The explanation was very clear and engaging, and I also like that we could try to apply what we learned during the workshop. I learnt not only the methods to debug faster but also the mindset to treat bugs as learning opportunities.”
“Very well-structured and good content. Well split up into different sections that are memorable and make sense. Great practical touch with the hands-on cases.”
“I really appreciated the workshop! What Juliano describes feels so obvious once he puts words to it, yet I haven't consciously thought about it before. He manages to concretise the hidden processes we all rely on in this field, and by articulating them, I feel more aware and intentional in how I debug and reason about ML problems.”
Anna Nylander, Smartr
“Great workshop that gave me new tools to debug machine learning models. The most interesting parts were how to prepare to minimize debugging time and the importance of reflection after solving a bug. Two things I take with me into my daily work.”
Robert Nyquist, Smartr
“Juliano identifies pain points we all encounter in ML work and provided practical strategies to avoid them. What resonated most with me was the mental shift: treating debugging not as firefighting, but as strategic information gathering. A game of maximizing insights while minimizing wasted effort.”
Fredrik Ring, Smartr
“The explanation was very clear and engaging, and I also like that we could try to apply what we learned during the workshop. I learnt not only the methods to debug faster but also the mindset to treat bugs as learning opportunities.”
“Very well-structured and good content. Well split up into different sections that are memorable and make sense. Great practical touch with the hands-on cases.”
Get in contact today to book a session.
Give your team the right mindset and tools to deal with ML systems in an effective way.