Carlos de la Cruz.
I build product software for AI workflows. My lanes are backend systems, applied AI, and full-stack AI product. Recent work includes a polymer-blend ML/RL system with TI Automotive, a career-exploration product used by millions of students at Xello, open education tools at Centennial's OER Lab, and vector search work at Exactera.
2,159 contributions in 2026
| Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | |||||||||||||||||||||||||||||||||||||||||
Recent writing
all essays →Selected work
№ 01 № 02
TI Automotive Polymer ML
Applied ML/RL system for proposing polymer blend recipes from OEM target specs, built around a forward model, physics features, uncertainty, and model-based search.
Python · PyTorch · scikit-learn · pandas · model-based RL
TryAgent CA
Multi-tenant agent-box platform with isolated customer runtimes, runtime-node allocation, lifecycle operations, and five live OVHcloud nodes generating roughly $120 MRR.
TypeScript · Effect TS · Go · OVHcloud