Sensed AI
sensed.aiHow Sensed AI Filtered 22 Engineers Down to 2 Top Hires
Sensed AI is an Earth Observation and AI company in Istanbul building scalable systems for urban and agricultural monitoring. Working with TeamCraft and Archi's Academy, they ran a multi-stage hackathon evaluating 22 full-stack engineers and identified the top two through team-based execution and AI-assisted workflow analysis. Read their official collaboration announcement on LinkedIn.
The Challenge: Senior Full-Stack Hiring for Production AI Systems
Sensed AI needed engineers who could maintain frontend interfaces and integrate complex backend AI pipelines. Theoretical knowledge of React or Python was easy to find. Proven production capability was not.
- Resumes claimed framework experience that didn't translate to scalable code
- Standard interviews provided no workflow visibility, no Git hygiene, no PR collaboration signal
- Team dynamics stayed invisible until after the hire
The TeamCraft Solution: Multi-Stage, Team-Based Hackathon
Sensed AI's hackathon ran in three structured phases:
1. Aptitude Baseline
A foundational quiz on core web technologies filtered out candidates lacking minimum technical baseline.
2. Team-Based Project Execution
Candidates worked in teams on real-world tasks. Across the multi-day sprint, the cohort resolved 22 tickets and submitted 22 pull requests in a simulated environment mirroring Sensed AI's actual engineering culture.
3. AI-Assisted Workflow Analysis
TeamCraft scored every candidate continuously across four dimensions:
| Evaluation Dimension | Cohort Avg Score |
|---|---|
| Code Intelligence (PR quality + architecture) | 60 |
| Project Execution (ticket completion + sprint velocity) | 82 |
| Technical Competency (framework integration) | 63 |
| Professional Behavior (collaboration + PR feedback) | 49 |
The data revealed a sharp split. Most candidates could execute tasks, but code intelligence and professional collaboration separated the top two from the rest.
Ready to see results like these?
Book a personalized demo and see how TeamCraft can transform your hiring.
Results
- 22 full-stack engineers evaluated in parallel
- 22 pull requests + 22 tickets completed across the cohort
- Top 2 candidates identified through data-driven scoring
- Workflow discipline + collaboration surfaced as decisive predictors, invisible in traditional interviews
For the full multi-stage scoring breakdown, see the Sensed AI case study on our blog.
Why It Worked
Sensed AI didn't just want engineers who could code. They needed engineers who could ship and collaborate inside a deep-tech engineering culture. The team-based hackathon revealed exactly that, before any offer was extended.
Frequently Asked Questions
Why did Sensed AI choose hackathon-based hiring over traditional interviews? Standard interviews can't show how engineers commit code, manage pull requests, or incorporate team feedback. All critical for Sensed AI's scalable AI systems work.
How were team dynamics evaluated? Candidates worked in groups, resolving real tickets together. TeamCraft scored their PR feedback patterns, communication, and how they handled blockers, producing the "Professional Behavior" score that ultimately separated finalists.
Can other AI companies replicate Sensed AI's approach? Yes. The multi-stage model (baseline quiz, team execution, AI workflow scoring) works for any engineering role where production readiness matters more than theoretical knowledge.