Mos Daniele

machine learning engineer · scientific computing

Bio

Machine Learning Engineer with a background in Computer Science and a Master's in HPC & AI. Taking ML systems from research to production across scientific computing, remote sensing, foundation models and agentic systems, currently CTO shipping geospatial ML to live clients. Drawn to deep tech: mission-driven work where ML accelerates scientific discovery.

Co-Founder of BorMos Tech. Most of my work can be found on Github and the Projects section.

Experience

2025.02 — now

Spatialise

Amsterdam

100k+ hectares mapped for clients at 10×10 m resolution.

2026.03 — now

CTO

Own tech operations for live clients, client prospecting and team planning. Build the platforms bringing our soil carbon services to EU and African farmers. Represent Spatialise tech in ESA's two-year IF-CSA programme.

→ ESA

2025.07 — 2026.03

Geospatial Deep Learning Researcher

Designed a novel relational GAT architecture fusing IBM's TerraMind geospatial foundation model with a mixture-of-experts head for soil organic carbon prediction, reaching state-of-the-art MAPE. Built its distributed multi-stage hyperparameter search and fine-tuned it on Snellius, the Dutch national supercomputer. Now in production.

→ paper

2025.02 — 2025.07

Machine Learning Engineer

Built the FastAPI inference services and Docker/GCP deployment, wired predictions into Retool.

2025 — now

BorMos Tech

Co-founder · part-time ML consulting · Remote

2026.07

Concurrences · Interim AI Lead (part-time)

Covered the AI lead for a month at a legal publisher building an antitrust-law AI assistant, bridging legal, editorial and engineering.

2025.12 — 2026.03

Facilioo · ML Engineer (contract)

Agentic triage for tenant complaints on a property-management platform: RAG plus a tool-calling agent, with LangSmith tracing and offline evals on sensitive tenant data.

2025.06 — 2025.11

Nexus Studios · ML Engineer (contract)

Built a computer vision pipeline for 3D mesh reconstruction of head geometries.

2024.04 — 2024.10

HZDR

ML Research Intern · Dresden

Built a parallel active learning pipeline on the Hemera HPC cluster for ML surrogate models of laser-plasma particle acceleration, the basis of my Master's thesis on model uncertainty.

2023.10 — 2024.09

Elbflorace Formula Student

Autonomous Systems Engineer · Dresden

Implemented LiDAR-based ground segmentation with multi-region RANSAC. Trained YOLO models for cone detection on an autonomous race car.

2021.09 — 2023.09

Tvarita

ML Engineer · Cluj-Napoca

Managed facial measurement data pipelines, trained lens segmentation models, and built an optics showcase application in C++ and Qt.

2021.07 — 2021.08

machinations.io

ML Intern · Cluj-Napoca

Built ML-based user prediction in a simulation environment. Developed MySQL data pipelines and trained PyTorch classification models.