ML and MLops on AWS

Da li ste se ikada zapitali šta se dešava kada model mašinskog učenja umre? Gde ide? Da li je zameniv? Možemo li, bez twitter-a, razotkriti njihove predrasude i greške? U ovom predavanju biće reči o životnom ciklusu modela mašinskog učenja, kako obezbediti njegovo zdravlje, dijagnostikovati kada je vreme da se izgradi novi i zameni stari model i kako da se, što je najvažnije, odupre porivu da se 'samo malko podesi' i bavi se obučavanje mlađe (i nadamo se sposobnije) zamene. Konačno, biće reči o preporučenim praksama za zamenu datih sistema bez prolivanja suza (lako ako ste robot bez duše), ili kako to učiniti na način koji dolikuje dobroćudnom kreatoru.
Bojan Miletic, predavac:

• Python developer preko 10 godina
• Voli jednoroge i kolače (spoiler)
• Sa AWS-om radi preko 6 godina
• Poseduje iskustvo u različitim ulogama, od frontenda do CEO. Okušao se čak i u dizajniranju pa su mu platili da se vrati kodiranju
• Voli da pomaže kompanijama da na valjan način razviju svoje ML proizvode
• Trenutno je CEO kompanije Softerrific

Softerrific is a fractional MLOps team that turns your ML algorithms into PoCs, MVPs, or full-fledged products that convincingly win over investors and bring real business value… And help you stand out in a saturated market like a unicorn at a pony show.

Our superpower is to strategically and sustainably #pythonize your AIdeas thanks to our passion for the ideation process. What can you expect from us?
📍 Developing ML pipelines
📍 Creating APIs for using ML models
📍 Developing PoCs and MVPs from ML models
📍 Deployment to the cloud and bigger-scale production
📍 Consulting for AI/data scientists on clean/reusable code writing practices
We’re a human-centered team that works with human-centered, middle-sized to enterprise-level AI companies.
🥇 Human-centered because every product or service should aim to push other people forward. To make their lives better. And you can’t have that if you don’t put HUMAN in the core of your product.