Co2Offset.ai — Forest Carbon SaaS

SaaS measuring forest carbon capture from satellite and ground data — cheap, fast and reliable, rewarding landowners for protecting and growing forests. Pombal, Portugal.

Highlights

• ML in production: Node.js microservices serving computer-vision models with real SLAs; PyTorch and scikit-learn training with evaluation and performance monitoring in production (torchvision, Pillow, scipy).
• Scale: designed and maintained 400+ containerised microservice endpoints and 200+ automation workflows (event-driven ingestion, CI/CD and ETL jobs) in Node.js and Docker with high-volume satellite-imagery ingestion; Kubernetes with auto-scaling and health probes; PostgreSQL / PostGIS, MongoDB and Redis configured and tuned, with ETL pipelines.
• Cloud, IaC & delivery: cloud-native, cloud-agnostic architecture on AWS with serverless components (Lambda, EC2, S3, RDS, EKS, IAM, R2); Terraform across multiple environments and Ansible for configuration and deployment; GitHub Actions CI/CD with PyTest; Vue.js / React frontends, Stripe payments; React Native / Expo mobile client shipped to the App Store and Play Store.
• European Space Agency: partnered directly with ESA on requirements, deliverables and geospatial data pipelines — translating complex geospatial and AI requirements into development specifications; Python microservices for ESA feeds.
• Leadership: ran feature prioritisation and sprint planning, coordinated application lifecycles, mentored junior engineers on MLOps and DevOps practice, and bridged technical teams and business stakeholders.
• Impact: led the pipeline audit with Bureau Veritas, securing the scientific-reliability certification that unlocked €2.5M of business; SOC 2 and ISO 27001 compliance; Three.js digital twin between client forests and the production engine.