~/about
About
FIELD FILE
IDENTIFICATIONSM-AGENT
NAME
Srdan Mijuskovic
ROLE
AI Product Manager & Systems Builder
BASE
Hamilton, ON · Canada
TIMEZONE
ET (UTC−5)
LANGUAGES
English (C1/C2) · Serbian (native)
FOCUS
AI Products · Consulting · Advisory
OPERATIONAL HISTORY
2021 → NOW
SR. PRODUCT MANAGER & DATA SCIENCE TEAM LEAD
Europe's leading classifieds platform · 5M+ monthly users
DEPLOYED
- →Customer support analytics · first-ever real-time KPIs (frustration, FCR)
- →AI Garage · internal LLM validation platform · PMs prototype without eng
- →HR chatbot · 80% adoption · 2,200+ queries
- →Recommender system · $200K grant · 6 months early
- →Anti-spam, content moderation, outreach chatbot
- →€500K big data platform · C-suite sign-off · 15+ engineers coordinated
IMPACT
5M+ users · €500K led · $200K grant
2016 → 2021
INDEPENDENT PRODUCT MANAGER
Premium beauty brand (international)
DEPLOYED
- →Full product lifecycle, from research to commercialisation
- →International manufacturing (Germany, Switzerland)
- →E-commerce build (Shopify, WooCommerce)
IMPACT
$500K+ portfolio launched
2016 → 2018
BUSINESS DEVELOPER & PRODUCT LEAD
Phiacademy (mobile learning platform)
DEPLOYED
- →Mobile learning courses · iOS/Android
- →Masterclass event platform with integrated ticketing
IMPACT
$1M+ event revenue · 10,000+ learners
TECHNICAL STACK
AI & Data
LLMs & Generative AISemantic SearchRAG & EmbeddingsRecommender SystemsNLPPrompt EngineeringModel Evaluation
Prototyping
Pythonn8nStreamlitFastAPINext.jsLovableCursor
PM Craft
0-to-1 DevelopmentMVP ValidationA/B TestingRoadmappingStakeholder ManagementBusiness Case Development
Tools
JiraNotionFigmaMixpanelGitHubSanityVercelMiroSuperset
EDUCATION
Information Technology School (College of Applied Studies)
Belgrade
2013–2015
Faculty of Philosophy, University of Belgrade
Belgrade
2011–2013
FIELD OBSERVATION
→ On building AI products
“I came to product management through philosophy and social sciences. That shapes how I think about users and failure modes. Not just optimising metrics, but asking why people behave the way they do.”
→ On prototyping
“Build small, validate fast, go big only when the signal is real. I've seen too many teams spend months shipping the wrong thing with full confidence.”
PRINCIPLES
01
Prototype before you spec
Writing a spec for something you haven't validated is expensive guessing. The prototype answers the question the spec assumes away.
02
Build for the real problem
The problem in the brief is rarely the real problem. The real one emerges from talking to users, not from reading the backlog.
03
Ship small, measure fast
Go big only when the signal is real. Confidence without evidence is just noise.
sm-agent · bash · Hamilton, ON
$ whoami
srdan-mijuskovic — AI Product Manager & Systems Builder
$ ls /shipped
support-analytics/ ai-garage/ hr-chatbot/
recommender-sys/ data-platform/ anti-spam/
$ git log --oneline -5
SPEC-001 Support analytics platform → real-time KPIs · Excel gone
SPEC-002 Recommender system → $200K grant · 6mo early
SPEC-003 HR chatbot → 80% adoption · 2,200+ queries
SPEC-004 Data platform €500K → 90%+ pipeline acceleration
SPEC-005 AI Garage → PMs prototype without eng
$ status --check
● AVAILABLE · Hamilton, ON · ET (UTC−5) · responds within 24h
$