Built a state-of-the-art automated home valuation system that beat 3,800+ teams.
Elite applied AI for prediction, markets, and real-world products.
Mindle builds high-accuracy machine learning systems for teams that need models to survive real data, real decisions, and real evaluation pressure.
Global top-tier competition record across ranking, forecasting, valuation, and search.
Ranked #1 in Canada and top 20 worldwide with 10 gold and 10 silver medals.
First inventor on a Zillow patent tied to core Zestimate modeling.
Award record
Proof from competitions where every error is measured.
The original Mindle site led with awards for a reason: public leaderboards make credibility unusually concrete. The redesigned site keeps that proof near the top and ties it to the work Mindle does now.
1st 
$1M Zillow Prize, carried forward from the original site
$1M Zillow Prize
3,800+ teams · 91 countries
Built a state-of-the-art automated home valuation system that won the world’s biggest machine learning competition.
Two Sigma Financial Modeling
2,000+ teams
Prize-winning market prediction work under limited hardware and computational time.
Optiver quantitative modeling
4,000+ competitors
Open-source quant ML work adopted and referenced by hedge fund practitioners.
WSDM / KKBox recommendation
1,000+ teams
Recommendation modeling for Asia’s leading music streaming service.
Home Depot search relevance
2,000+ teams
Search relevance modeling to improve product discovery and shopping experience.
Rossmann sales forecasting
3,738 data scientists
Daily store-sales forecasting across 1,115 drug stores using historical, store, product, and third-party data.
Invited speaker
Nima Shahbazi, public speaker at top-tier AI conferences.
Invited speaker at leading AI and applied learning venues including KDD, Sigmod, TMLS, AIC Ontario, Georgian programs, ReWork, and RBC Disruptors.
KDD
Canada AI Day speaker
Speaker listing and York University coverage for the Canada AI Day trust panel.
Speaker page
AIC Ontario
Evolving with AI 2026
AIC Ontario speaker page and public keynote announcement.
Speaker page
Georgian
TLIR reinforcement learning
Applied learning session and public follow-up for Georgian programming.
Session page
ReWork
Machine Learning Summit Montreal
Presented on reducing demand forecast error with deep learning.
Event page
RBC Disruptors
Public AI and ML panel
RBC coverage of AI entrepreneurs speaking at its Disruptors forum.
RBC storyHedge fund strength
Quantitative ML shaped by market-prediction pressure.
Mindle's finance work is grounded in public, competitive evidence: Two Sigma, Optiver, and open-source quant ML work referenced by hedge fund practitioners.
Two Sigma quantitative modeling
Prize-winning work in market prediction under constrained compute, focused on signal discovery and robust feature evaluation.
View sourceOptiver market prediction
Top-10 placement among 4,000+ competitors, with open-source quant ML work referenced by hedge fund practitioners.
View sourceResearch workflow discipline
Competition-grade validation habits: careful leakage control, feature audits, fast experimentation, and benchmark-driven model improvement.
View sourceServices
From uncertain data to deployable model advantage.
Engagements stay concrete: define the decision, validate the data, build the model, measure the lift, and leave the team stronger.
Hedge fund and quantitative finance ML
Support alpha research, market prediction, signal discovery, feature engineering, model validation, and production research workflows.
Applied prediction systemsDesign and improve models for valuation, forecasting, ranking, recommendation, search relevance, pricing, and operational decision systems.
AI product discoveryTurn ambiguous business problems into low-risk AI product proposals with feasibility, data needs, roadmap, timeline, and cost clarity.
Model performance rescueDiagnose underperforming ML systems, repair evaluation loops, improve benchmarks, and reduce error with disciplined experimentation.
Products
Mindle also builds AI products.
Olva is the first product featured in the Mindle portfolio: a private, desktop-first meeting companion for live answers and post-meeting memory.
Developed by Mindle
Olva
Invisible real-time meeting assistant for professionals.
- Live transcription and useful cues
- Automatic question detection
- AI Boost and document-aware answers
Selected achievements
A record across hard prediction problems.
The through-line is not one domain. It is disciplined modeling under measurable pressure.
Real estate valuation
- Winner of the $1M Zillow Prize
- Automated home valuation model
- Zestimate patent holder
Quantitative finance
- Two Sigma prize winner
- Optiver top-10 placement
- Open-source quant ML adoption
Search and recommendation
- WSDM/KKBox recommendation challenge
- Home Depot search relevance
- Ranking and relevance modeling
Forecasting and pricing
- Rossmann sales forecasting
- Mercari price suggestion
- Demand and price prediction
Trustworthy AI
- $1M Leaders Prize finalist
- AI systems for misinformation
- Public speaking on trustworthy AI
Press and public work
Recognized by technical, business, and AI communities.
GeekWire · Jan 30, 2019 Meet the 'Zillow Prize' winners who get $1M and bragging rightsVentureBeat · Jan 30, 2019 Zillow awards $1 million to team that reduced home valuation algorithm errorNVIDIA Developer Blog · Jan 31, 2019 NVIDIA GPUs help developers score $1 million prize for improving Zillow's ZestimateKaggle Blog · May 25, 2017 Two Sigma Financial Modeling Challenge winner interview
Start a conversation