Cloud Simulation Job Platform
An Azure job platform for long-running physics simulations: FastAPI job API, queue-triggered workers that scale to zero, React + TypeScript front end
Pablo Arango
I take ML models, LLM pipelines and simulation platforms from prototype to production. ML and LLM systems, prototype to production.
Architect turned engineer: I build the tools I used to wait on, turning hour-long physics simulations into answers in under 10 seconds.
5 → 50+ ML usersML users
20–100× faster than simulationfaster
500+ design runs / monthruns / month
7 yrs professional PythonPython
Selected work
Production systems & projects
Graph Neural Network for Overheating
TM59 and Part O compliance in under 10 seconds, ~100x faster than IES VE
Vision-LLM Event Extraction Pipeline
A scheduled pipeline that reads tour operators' Instagram carousels with a vision model and publishes a searchable outings site
Property Valuation Engine
Fair asking price per m², with a calibrated interval and the comparables behind it, for every home listed for sale in the Medellín metro area
LiDAR2Building Geospatial Engine
3D context geometry (extruded buildings, terrain, streets, trees) from a lat/lng for anywhere in England
CNN-Based Daylight Autonomy Prediction
Real-time daylight autonomy at ~3% MAE, ~20x faster than Radiance
Profile
From simulation to production
7 years of Python in building-performance simulation (Transsolar, Hoare Lea). On contract to Hoare Lea I took its ML services to production, used by 50+ engineers, and I'm architecting its cloud platform for long-running simulations. I build agent-native: I write the specs and acceptance criteria, and I own the design, code review and verification.
Role
CurrentSenior AI Engineer · ML & Simulation Platforms
EngagementHoare Lea (contract, remote)
Based inMedellín, Colombia (UTC-5)
Credentials
EducationArchitect, Summa Cum Laude, UNAL Colombia
FellowshipTranssolar Academy 7, Stuttgart
Stack
BackendPython · FastAPI · Docker · Azure · Bicep
AI & MLLLM agents · evals · MCP · PyTorch · Azure ML
Front endReact · TypeScript · Vue.js · Streamlit
GeometryRhino · Grasshopper · rhino3dm
SimulationRadiance · EnergyPlus · TAS · OpenFOAM
Services
What I offer
AI Agents & LLM Pipelines
Agents and LLM pipelines that classify, extract and summarize, each with an eval and a cost ceiling.
Learn more →ML Models to Production
From notebook to a service engineers use daily: data pipelines, FastAPI serving, cloud deployment.
Learn more →Simulation Compute Platforms
Cloud job platforms for long-running workloads: job APIs, workers that scale to zero, IaC and tracing.
Learn more →Get in touch
Get in touch about engineering, ML or simulation work.
Open to senior roles, full-time or contract, and to scoped contract work. Remote from Medellín (UTC-5), with full US Eastern overlap and hours that work for UK and European teams.
Get in touch