Train Matricx

BACKGROUND

Introduce your business and what you do there.

I’m a product manager at a Product Development Company. We help startups, SaaS companies, and growing businesses turn ideas into practical products through thoughtful design, reliable software, and AI that belongs in the real world.

OPPORTUNITY / CHALLENGE

What challenge were you trying to address with Train Matricx?

We needed help producing box scores from basketball game footage at a scale our own team couldn’t cover. Our platform depends on verified stats, but many amateur games have video and no statistician, so those players never show up in our data.

SOLUTION

What was the scope of their involvement?

Train Matrix handled video-based box score creation for amateur basketball games. Their team watched full game footage and recorded player-level stats for each game, including points, field goals, 3-pointers made/attempted, free throws, rebounds, assists, steals, blocks, turnovers, fouls, and minutes played. Every game went through a two-layer QC process before delivery.The larger goal was to build a model that detects and generates box scores for amateur basketball players. To train and validate it, we needed a large set of an-verified box scores. Train Matrix completed 745 basketball games for us with 90% accuracy, using a two-layer QC process to review every game.

What is the team composition?

We worked with 2–3 teammates from Train Matricx.

How did you come to work with Train Matricx?

We found them through an online search. We chose them because their pricing fit our budget, they offered good value for the cost, and their company values aligned with ours.

How much have you invested with them?

We spent $20,000.

What is the status of this engagement?

We worked together from July-September 2026.

RESULTS & FEEDBACK

What evidence can you share that demonstrates the impact of the engagement?

Train Matrix delivered 745 games with 90% accuracy and a 24-hour turnaround time. That gave us the verified dataset we needed for our amateur box score detection model. The 745-game verified dataset gave us the training and benchmark data to build our amateur box score detection model, and it moved us from no scalable coverage of amateur games to a working path to automate it.

How did Train Matricx perform from a project management standpoint?

Train Matrix delivered on time. We communicated via virtual meetings, email, and messaging apps.

What did you find most impressive about them?

What stood out most was the combination of scale, speed, and consistency. Train Matrix completed 745 basketball games for us, each with a 24-hour turnaround, and every game went through a two-layer QC process. That level of reliability at that volume is hard to find. Many vendors can do one or the other, but holding a steady standard across hundreds of games is what made the dataset usable for training our model.They also understood basketball. Their annotators knew the game, so we didn’t have to explain what a rebound or a steal looks like.

Are there any areas they could improve?

No, there weren’t any.

Do you have any advice for potential customers?

I’d recommend them to any basketball data or AI team that needs reliable stat-by-video work at scale.