Kpler · 2021–2025 · Staff Data Engineer — founder & manager, MLOps squad
MLOps platform & squad, from scratch
Founded and managed Kpler's MLOps squad and built its self-service platform for training (GPU), deploying and serving models in production — the foundation that industrialized AI/data use cases while the company scaled from ~100 to 750+ people and crossed $100M ARR.
- Self-service GPU training, deployment and inference on AWS EKS with Terraform, GitOps and CI/CD.
- In-house framework of shared modules (helpers, observability, alerting) with standardized SDLC.
- "You build it, you run it": data scientists owned their use cases in production, on-call included.
- AWS
- EKS
- Terraform
- Airflow
- GitHub Actions
- Helm
- Aurora PostgreSQL
- Elasticsearch
- MLOps