The challenge
A technology company building a computer vision system for customer analytics needed technical leadership. Their previous AI lead had departed, leaving the project without:
- Clear technical direction or roadmap
- Realistic timeline estimates
- Any training data – the dataset didn’t even exist yet
- Documentation for data preparation and model development
- Validated hardware specifications
The team needed someone who could quickly assess the project’s technical requirements, deliver a viable implementation strategy, and build working systems from the ground up.
My approach
Collaboration with stakeholders
During the entire engagement, I collaborated closely with the CEO and the rest of the team, where I shared my ML knowledge with them so they could better understand the scope of the project.
Technical assessment and planning
I started by understanding the full scope: a distributed computer vision system for tracking customer behavior and engagement metrics. With no existing dataset, I had to plan both the data collection pipeline and the ML development:
- AI roadmap – defined the development sequence and technical milestones
- Implementation timeline – realistic estimates that accounted for different phases of the project
- Data annotation guide – documented requirements for training data collection and annotation
- Hardware validation report – tested and validated the existing hardware for ML workloads
Hands-on development
Beyond strategy, I built working systems:
- Developed proof-of-concept code demonstrating core system functionalities
- Managed and assigned tasks to a junior developer
- Wrote a codebase that formed the foundation for continued development
Leadership transition
When it came time to transition the project, I:
- Participated in hiring process to find the right replacement AI lead
- Ensured all strategic documentation was in place for continuity
- Left the project with clear technical direction and a viable path from zero data to working system
Results
- Delivered complete technical foundation including roadmap, timeline, and working code
- Established realistic expectations with stakeholders about project complexity and timelines
- Enabled smooth leadership transition – the incoming AI Lead had clear direction to build on
Key insight
When AI projects lose technical leadership, the path forward often becomes unclear. This engagement demonstrated the value of bringing in experienced ML expertise to quickly assess the situation, document a realistic strategy, and build working systems that enable the team to continue successfully – even when starting with no data at all.