AI/ML Engineer Volunteer
Must-have:BackendDataAI
AI/ML Engineer (Volunteer): Autonomous Route Intelligence
Role summary: You will build the intelligence layer of Yspace's Autonomous Route Intelligence. The product helps logistics companies and riders choose faster, more reliable routes, using provider routing data plus Yspace's own learning.
Roles and Responsibilities
- Own the ML layer of Route Intelligence. Design how machine learning sits on top of the routing engine. The routing engine computes the route and ETA, and your models learn the gap between that ETA and what actually happened.
- Build the data foundation with the backend and data engineers. Define what gets logged: every route request, every recommendation, and the actual arrival outcome. You are responsible for data quality, schemas and labeling. This includes the feedback path where companies send back real trip results.
- Develop and evaluate ETA prediction models. Start with strong baselines such as gradient boosting. Measure error (MAE), broken down by time of day, vehicle type and area. Move to deep learning only when the data and results justify it.
- Build route scoring and reliability ranking. Compare candidate routes (Recommended, Alternative A, Alternative B) and rank them. Begin with transparent, rule-based and statistical scoring, then introduce learned scoring as data grows.
- Enable enterprise-specific preferences. Some companies value reliability, some speed, some fuel efficiency. Represent these preferences so that scoring adapts to each business.
- Introduce risk and anomaly detection responsibly. Build delay-risk and unusual-travel-time detection only once there is enough data to support them. Never present predictions as intelligence that hasn't been validated.
- Make recommendations explainable.** Support plain-language explanations such as "Route B is about 8 minutes faster because of current traffic." Every claim must trace back to real signals and measured data.
- Deploy and operate models in production. Serve low-latency predictions through the API. Build fallback behavior for when a model or provider is unavailable. Handle model versioning, monitoring, drift detection and retraining.
- Develop local logistics intelligence for Nigeria. Test provider map and traffic quality in Lagos, Abuja and other target cities. Handle messy addresses and local road behavior. Where provider data is thin, design ways to collect Yspace's own data.
- Collaborate, document and protect data. Work with backend, mobile (SDK) and product teams. Write clear model documentation and evaluation reports. Respect data ownership and privacy: customer data stays separated by organization, and limitations are communicated honestly.
<