Denis Nedeljkovic
I am an AI/ML Developer and Data Scientist based in Toronto, specializing in building scalable AI pipelines and optimizing predictive models for urban safety.
About
I am an AI/ML Developer and Data Scientist based in Toronto, specializing in building scalable AI pipelines and optimizing predictive models for urban safety. With expertise in Python, microservices, and hybrid infrastructure, I actively collaborate with leadership teams to streamline product delivery and innovate technical solutions. I also have hands-on experience developing SaaS platforms and 3D AI-to-print workflows for manufacturing.
My work experience spans research and applied machine learning projects that significantly enhance computational efficiencies and model performance. I am passionate about leveraging AI agents, cloud computing, and explainable AI techniques to solve real-world problems while ensuring security and cost-effectiveness. I continuously aim to create impactful technology that benefits both industry and government sectors.
- 3 years' experience
- Joined September 2025
Languages
- AfarAdvanced
- JavaneseIntermediate
Industry experience
- Government
- Manufacturing
- Software & Internet
- Education
Work experience
Built scalable AI pipelines integrating GPT-based agents with internal tools, reducing task overhead for leadership by over 60%. Engineered a full 3D AI-to-print workflow combining 2D-to-3D conversion, mesh repair, and export automation for manufacturing. Designed hybrid infrastructure using on-prem GPU servers and cloud instances to optimize compute cost and security. Led internal MVPs for AI agents, design automation, and executive backup systems, streamlining product planning and delivery. Coordinated technical strategy with CEO and CTO, contributing to architecture and system decisions.
Developed predictive models optimizing urban safety through strategic sensor placement, significantly improving accuracy and computational efficiency. Integrated and preprocessed diverse datasets with extensive feature engineering, reducing computational requirements by about 55% while maintaining high model performance. Enhanced processing speed using load balancing, multithreading, and parallel computing, cutting computational time from roughly 672 hours to 65 hours. Designed a scalable, cloud-based SaaS solution with a user-friendly GUI to help government and law enforcement visualize traffic hotspots for proactive safety planning.
Education
Hire Denis Nedeljkovic today
Post a job and invite Denis, or contact them directly. No hidden commission.