Bojan Savić

I started in L&D the way a lot of people do — teaching, freelancing, building courses for anyone who would pay for them. The unusual turn came early: I did not stop at designing the courses. I started building the tooling around them. An improvised LMS in Articulate Storyline 2. A Google Docs add-on that brought AI writing inside the document. A bespoke ROI tracking system on Wrike and Tableau. If a tool was missing, I made one.

That instinct took me across a strange list of employers. Humanitarian operations at the ICRC, where I led a cross-functional team of ten and shipped a field-deployable learning product for staff in active conflict zones. Government consulting with GOPA / GIZ, rebuilding a ministry's internal learning infrastructure. NGO work with WWF, Catalyst Balkans, and Nordeus Foundation. An EdTech startup where the CEO was my direct report. Most recently Booking.com, where I lead learning at enterprise scale — with a six-month chapter as the caretaker team lead.

The through-line is the same at every stop: measurable behaviour change beats pretty artifacts. If a course did not move the needle on real operational signals — CSAT, quality scores, incident rates, honest Monday-morning behaviour — it was training theatre, and I would rather rebuild it than ship it.

The receipts, for what it's worth: a strong track record in recent performance cycles, an acting team-lead appointment on the back of that, and a custom learning-impact metric adopted by stakeholders and built into my team's dashboard. I lead with those reluctantly — they are lagging indicators of having the argument right, not the argument itself.

On the side, I build what the day job reminds me is missing. Scenario Studio, a Godot-based serious-game authoring tool for instructional designers who hit the ceiling of Storyline and H5P. Content Factory, a code-first video pipeline that compiles an idea into a published clip in a single command. Suprender, a small game studio I co-run — currently developing Lootbane, an ARPG/roguelite. Separately, weekend Godot jam entries under QMachines and ad-hoc teams. Plus a rotating set of smaller prototypes that keep me honest about where AI is useful in L&D, and where it is still snake oil.

I write because my own mistakes are the cheapest tuition I can offer other people. The tone is collegial skeptic: the work goes first, the claims come after the evidence, and the receipts are public where I can make them public.