AIRINA's institutional partners — banks, MFIs, telecoms, public-sector engineering — employ thousands of engineers whose work could be AI-augmented but who lack the training. This program is for them.
Five days, hands-on, taught in Cotonou or live online — designing AI-enabled engineering systems, working with pretrained models, deployment patterns including on-device and edge AI, and applied generative AI for engineering workflow. The Cotonou edition adds a focus on what AI looks like under West African infrastructure constraints: intermittent connectivity, low-power compute, mission-relevant problems in transport, energy, and water.
Program Overview
Five days, full-day sessions. Mornings on the concept and the architecture; afternoons in the notebook, on real engineering datasets. Pretrained models throughout — the program is about using AI as an engineering primitive, not about training new architectures from scratch. The final day moves into applied generative AI for engineering workflow and a small graded project.
Program structure
- Day 1. Foundations — when is AI the right answer for an engineering problem and when is it not. Working with pretrained models, the Hugging Face ecosystem, model selection criteria.
- Day 2. Computer vision for engineering — defect detection, object localization, time-series imaging. Hands-on with pretrained vision models on real engineering data.
- Day 3. Time-series and signals — anomaly detection, predictive maintenance, sensor data fusion. Hands-on lab.
- Day 4. Deployment under infrastructure constraints — on-device inference, model quantization and distillation, edge AI, the realities of intermittent connectivity in West African deployment contexts.
- Day 5. Applied generative AI for engineering workflow — LLMs as design assistants, documentation generators, code copilots; safety and verification in AI-assisted engineering. Final defended project.
Certificate
Successful completion — attendance plus the Day-5 defended project — earns an AIRINA AI for Engineers certificate, graded.
Learning Outcomes
By the end of the program, participants will be able to:
- Recognize when an engineering problem is well-posed for an AI approach — and when AI is the wrong tool.
- Select a pretrained model for a given engineering task and adapt it (transfer learning, fine-tuning, prompting) to local data.
- Build a working computer-vision pipeline for an engineering use-case (defect detection, monitoring, asset inspection) end to end.
- Build a working time-series anomaly-detection or predictive-maintenance pipeline on sensor data.
- Deploy a model under infrastructure constraints — pick on-device vs cloud, quantize for edge, design for intermittent connectivity.
- Use LLMs as an engineering workflow tool — documentation, code, design review — with explicit verification practices, not blind trust.
- Defend a small applied-AI engineering project end to end: problem framing, model choice, evaluation, deployment plan.
Program curriculum (5 days)
- Where AI fits in an engineering pipeline (and where it doesn't)
- The Hugging Face ecosystem — model hub, datasets, the pipeline API
- Model selection criteria for engineering tasks
- Hands-on: load a pretrained model, run inference on real data, evaluate honestly
- Defect detection, object localization, segmentation — from pretrained backbones
- Time-series imaging — spectrograms, scalograms, the visualization trick
- Adapting models to local data — fine-tuning vs feature-extraction
- Hands-on: vision pipeline on a real asset-inspection dataset
- Anomaly detection — statistical, ML-based, and hybrid approaches
- Predictive maintenance — framing, evaluation, the cost-of-failure asymmetry
- Sensor data fusion — the principles, the gotchas
- Hands-on: build and evaluate a predictive-maintenance pipeline
- On-device vs cloud — cost, latency, privacy, connectivity
- Model quantization and distillation — making a model fit the device
- Edge AI hardware — what's available, what's affordable, what fails
- Designing for intermittent connectivity in West African deployment contexts
- LLMs as design assistants, documentation generators, code copilots
- Safety and verification when using AI in engineering deliverables — what you cannot trust the model to get right
Final defended project: take an engineering problem from your own work, build a working AI-augmented solution, defend it in front of the cohort and AIRINA faculty. Written critique returned within a week.
Who Should Attend
This program is for working engineers who want to add AI to their toolkit — with hands-on practice on real engineering data, under the infrastructure constraints AIRINA's institutional partners actually operate under.
- Working engineers in mechanical, electrical, civil, or software fields wanting to add AI capability.
- Engineering team leads scoping AI integration for their team's work.
- Recent engineering graduates building applied-AI portfolios as part of an early career.
- Institutional engineers in West Africa working under infrastructure constraints (intermittent connectivity, low-power compute).
- Technical staff at banks / telecoms / utilities whose engineering work could be AI-augmented but who don't yet have the training.
Prerequisites
- Programming. Working knowledge of at least one programming language. Python preferred; if not Python, you should be able to read it.
- Engineering background. Either a formal engineering degree or several years of engineering practice. The program assumes you understand the kind of problems engineers solve.
- No prior AI / ML required. The program teaches working AI as an engineering tool from the ground up — what you bring is the engineering side.
Selection
For oversubscribed cohorts, applicants submit a one-paragraph description of an engineering problem from their current work that they want to attempt as the Day-5 project. Selection priority is given to applicants whose problem is most likely to produce a useful working artifact, and to underrepresented groups in technical fields.
Brochure
The detailed program brochure (PDF, EN/FR) is sent on request — including day-by-day curriculum, the industry-collaborator profile for the active cohort, dataset list for the hands-on labs, and the cohort calendar.
To receive the current brochure, write to contact@airina.africa with "AI for Engineers Cotonou — brochure request" in the subject.