FITIZENS
AI-powered movement analysis platform for functional fitness.
Exploring strategic acquisition and acqui-hire opportunities.
What You're Acquiring
LLM-as-a-Judge Video Analysis
Production-deployed multi-stage pipeline that analyzes exercise video, scores movement quality per rep, and generates coaching feedback. LLM-agnostic (currently Gemini, easily portable).
Exercise Catalog (529+)
Biomechanical rubrics, muscle activation percentages, bilingual instructions (EN/ES), common mistakes, and quality criteria for each exercise.
Labeled Sensor Dataset (40K+ reps)
9-axis IMU data across 130 exercises, recorded by 16-18 professional trainers. Consent cleared. Not replicable with AI tools.
Evaluation Pipeline (900+ videos)
Human-annotated ground truth for LLM quality assurance. Automatic comparison across prompt versions and model updates.
Published Mobile App
Flutter, iOS + Android. Seven feature modules. On-device video processing. Subscription infrastructure (Stripe, Apple IAP, Google Play).
30 ML Models
Temporal Convolutional Networks for exercise detection from IMU data. Quantized to TFLite, on-device inference under 5ms.
Full Backend Infrastructure
Firebase Cloud Functions (Python), Firestore, Cloud Storage, authentication, subscription billing (Stripe, Apple IAP, Google Play), email automation, analytics.
The Team
Both founders available for acqui-hire with retention.
Victor Gonzalez Pacheco
Co-Founder & CEO
- PhD in Robotics & AI
- Executive MBA, IE Business School
- 20 years building tech in AI & robotics
- Adjunct Professor, IE University (AI & Technology, undergrad to MBA)
Production expertise
- ● AI product strategy and development
- ● Full-stack agentic AI engineering
- ● Flutter, Python, Firebase, frontend development
Daniel Sierra
Co-Founder & CTO
- 11+ years shipping AI and data systems to production
- Former Lead Data Scientist at Telefonica Tech (enterprise AI)
- Adjunct Professor, IE University (Python & AI, 300+ students)
Production expertise
- ● AI product engineering: from prototype to production at scale
- ● Full-stack agentic AI: Python, Flutter, C++, Firebase, Cloud
- ● End-to-end ML pipelines: data acquisition, labeling, training, eval, deployment
- ● On-device ML and sensor systems: model quantization, TFLite, IMU fusion, dataset curation
Strategic Fit
For a coaching platform
Plug-in AI video analysis for 529+ exercises. Instant differentiation vs competitors. White-label the analysis pipeline into your existing app.
For a wearable company
40K+ reps of labeled sensor data and trained models for 130 exercises. Plus a pipeline to scale data acquisition automatically. Extend your value proposition into strength training.
For any AI company
Two senior engineers with production experience in agentic systems, LLM evals, and ML pipelines. Hard to hire on the open market in 2026.
The Product
Record a video of your exercise. The AI counts reps, scores each one, detects technique issues, and gives specific coaching feedback. 529+ exercises supported.
Per-rep feedback overlay with issue detection and coaching cues
Overall set score with quality level, rep count, and no-rep detection
Actionable improvement areas with frequency tags and specific corrections
By The Numbers
Why Now
No direct competitor in AI video analysis for functional fitness movements
Legacy fitness platforms actively acquiring AI-native startups: Strava acquired Runna, MyFitnessPal acquired Cal AI, SWORD acquired Kaia Health
Wearable companies (Whoop, Garmin) seeking movement detection capabilities beyond step counting
Coaching platforms lacking video analysis technology as a differentiator
The cost of replicating this in-house: 2 senior engineers x 4 years = 8 person-years at market rates
Next Step
We are happy to share a detailed technical package, data room access, or schedule an introductory call.
Victor Gonzalez Pacheco
Co-founder & CEO
victor@fitizens.io
