Olle Oltorp
M.Sc. Machine Learning
Industrial Engineering and Management, KTH
Education
M.Sc. Industrial Engineering and Management
2025–2027 (expected)KTH Royal Institute of Technology · Stockholm
B.Sc. Industrial Engineering and Management
2022–2025KTH · Stockholm
- Implemented the collaborative filtering system: a cross-domain latent factor model over a joint article–newsletter matrix, transferring implicit article read-time signal to predict newsletter engagement.
- Worked directly with the client, a Swedish media company, scoping what their data could actually support and assessing whether their article personalisation was worth extending to newsletters.
Upper Secondary
2019–2022ProCivitas Privata Gymnasium Karlberg · Stockholm
Experience
Research Assistant
Apr–Aug 2026KTH, Dept. of Industrial Economics and Management · Stockholm
AI adoption
WASP-HS- Reviewed and coded 33 empirical studies of organisations adopting AI into a common framework, synthesising the recurring barriers to adoption and how firms addressed them.
Sustainable food
Digital Futures- Ran the spend analysis end to end, turning a year of KTH's SEK 11.3M catering data into the university's first measurement of how far its own sustainability standard had actually reached.
- Identified key stakeholders, designed the interview guide, conducted the interviews and sole-authored the report, which set out why uptake had stalled and what would change it.
Admissions Data Assistant
Feb–Mar 2026KTH, Dept. of Industrial Economics and Management · Stockholm
- Cleaned and validated confidential applicant data for KTH's international master's admissions round, recalculating grade averages and verifying university rankings.
Property Maintenance Worker
2022–2025SKB (Stockholms Kooperativa Bostadsförening) · Stockholm
- Re-hired four consecutive summers.
Upcoming
Teaching Assistant
2026–2027KTH, School of Electrical Engineering and Computer Science · Stockholm
- Appointed teaching assistant for two second-cycle courses:
Projects
The Diamond Index
2026– · LiveA discovery layer over a digital asset marketplace that cannot be sorted by the attribute that determines value: aggregates the community grading standards defining rarity and screens every live listing against them.
Workout Tracker
2026– · LivePhone-first workout logger and training analysis. Originally built for personal use, since developed into a product with user accounts and subscription billing.
Across both projects
- Sole developer on both, end to end: product and interface design, frontend, backend, data layer, containerised deployment, DNS and the ops that keep them running.
- Both developed under a self-designed multi-agent development framework: work routed across models by cost and capability, with per-role scoping, token budgets and a mandatory independent review pass on every change.
Skills
- Programming
- Python
- Java
- SQL
- Libraries & tools
- PyTorch
- Scikit-learn
- Pandas
- Polars
- Git
- Docker
- Excel
- PowerPoint
- AI workflows
- Multi-agent workflow design
- Model-role delegation (coordinator, implementer, reviewer)
- Precise task decomposition
- Token and cost optimisation
- Recommenders
- Collaborative filtering
- Content-based and hybrid systems
- Matrix factorisation
- Implicit feedback
- Indexing and ranking
- Deep learning
- Neural networks and deep architectures
- Language models and embeddings
- Classical ML
- Regularised regression
- Tree ensembles
- Support vector machines
- Hidden Markov models
- Research
- Data preparation
- Semi-structured interviewing
- Qualitative coding and thematic analysis
- Business
- Corporate finance
- Management control
- Operations strategy
- Technology shifts and innovation
- Spoken
- Swedish (native)
- English (fluent)