**Senior Computer Vision Engineer – Human Pose & Biomechanics
Location:** Remote (Global) Type: Full-Time or Contract Company: Texas Sports Academy
About the Role
We are building an AI-powered training app with elite volleyball leadership (including University of Texas coaching staff).
The goal:
An app that watches an athlete perform drills and provides intelligent, biomechanically sound feedback on their form.
This is applied AI at a high level — not research for research’s sake.
We need a senior engineer who understands:
- Human pose estimation
- Temporal modeling
- Video pipelines
- Applied deep learning
- Biomechanics-driven feature extraction
What You’ll Build
- Video ingestion pipeline
- Pose estimation integration
- Joint angle calculation systems
- Movement scoring models
- Feedback generation engine
- Scalable architecture for mobile + backend integration
You will be a foundational technical architect of this product.
Required Experience
- 5+ years in computer vision or applied ML
- Strong Python skills
- Experience with human pose estimation frameworks (MediaPipe, OpenPose, MoveNet, BlazePose, HRNet, etc.)
- Experience processing and analyzing video data
- Deep learning experience (PyTorch or TensorFlow)
- Experience designing production ML systems
You must understand:
- Joint angle computation
- Temporal smoothing
- Movement sequence modeling
- Feature extraction from keypoints
- Real-world model limitations
Bonus Points
- Athletic or sports background
- Experience building mobile ML systems
- Experience deploying ML to edge devices
- Experience with 3D pose estimation
- Startup experience
What Success Looks Like
Within 90 days:
- Working squat grading prototype
- Clear pose-based feature extraction framework
- Reliable joint angle calculations
- Movement scoring logic
- Architecture roadmap for volleyball drill analysis
Take-Home Evaluation
You will build a minimal squat grading app:
Requirements:
- User uploads squat video
- Extract keypoints
- Calculate:
- Knee angle
- Hip angle
- Depth
- Back angle
- Output:
- Score (1–10)
- 3 actionable improvement suggestions
Deliverables:
- GitHub repo
- README explaining:
- Model choice
- Tradeoffs
- Scaling plan
- Limitations
Time expectation: 6–8 hours.
Compensation
Competitive. Open to global talent. Contract or full-time available.
We are not looking for someone who has “experimented” with pose estimation.
We are looking for someone who can build a real product.
Job Type: Full-time
Pay: $250,000.00 - $300,000.00 per year
Work Location: Remote