Our Approach

We are not your typical consultants - our founder, Filip Michalsky, been heads down passionately implementing AI algorithms since 2017. Our goal is to help you navigate the minefield of technical AI projects and reach successful outcomes.

Open source contributions

We are active in the OSS community:

SalesGPT - Context-aware AI Sales Agent to automate sales outreach.

Algorithms Development

Filip has successfully completed many large scale AI projects in his data science career:

• Built more AI training and inference pipelines including CI/CD than he can count.
• Driving multi-arm bandit recommendation models at Fidelity, powering online recommendations for 35M+ retail brokerage customers.
• Developed multi-objective recommendation systems with custom loss function, driving 50% CTR and 20% take action rate on a Pareto frontier across millions of users.
• Trained and fine-tuned GPT models since GPT-1 release.
• Developed "Zestimate for Commercial Real Estate" using a mixture of gradient boosted trees and robust data extraction pipelines, covering more than 25% of US metropolitan statistical areas (MSAs).

AI Research

Filip has published research in top AI conferences:

• Fairness Evaluation for Uplift Modeling in the Absence of Ground Truth (ICMLA 2024)
• Surrogate Ground Truth Generation to Enhance Binary Fairness Evaluation in Uplift Modeling (ICMLA 2021)
• Fairness Criteria for Face Recognition Applications (AAAI 2019)

Why do we do this?

We are passionate about "applied AI" which means solving real world problems and building products people love.

We are going to work with you deep in the trenches to get value out of you working with us.