CV
The employment side of the ledger: enterprise AI/ML pre-sales at Dassault Systèmes, and the working method behind the projects on the rest of this site. For the projects themselves, start at the lab.
WHAT I DO AT DASSAULT.
Technical pre-sales for APAC enterprise sales cycles — from first discovery workshop to signed deal.
Industry Process Consultant — AI/ML Solutions | Technical Pre-Sales
AI-focused pre-sales for enterprise customers across APAC: AI/ML demos, POC architecture, customer workshops, RFx contributions, and executive technical presentations. Builds RAG/LLM knowledge-assistant prototypes and AI-powered MODSIM demonstrations.
Industry Process Consultant — Tech Sales, Asia Pacific South
Technical pre-sales across data platforms, supply-chain optimisation, and simulation on the 3DEXPERIENCE platform. Built a natural-disaster-risk simulation (Python, geospatial data, ML, cellular automata) used as a flagship customer demonstration. Level 3 certified in DELMIA Quintiq Workforce & Macro Planner.
Workshops & requirements mapping
Discovery sessions with pre-sales, product (brand) teams, and regional specialists — across Australia and globally with Paris — scoping customer problems across mining, energy, automotive, infrastructure, and defence.
Custom demos that close deals
Tailored demos and POC environments for enterprise prospects. Not slide decks — working software configured to their data, their workflows, their language.
Platform design & integration planning
How the platform fits existing tech stacks — API integrations, data pipelines, deployment models. Across simulation, data science, and supply chain.
Technical responses & solution proposals
Leading the technical side of RFPs and RFIs — fit/gap analysis, scope definition, data requirements, solution architecture, and presenting to customer evaluation panels.
An end-to-end maritime-monitoring prototype: Sentinel-1 SAR + Sentinel-2 multispectral imagery processed through fine-tuned Earth-observation foundation models (TerraMind, Prithvi-EO 2.0 600M), a custom U-Net, and specialist marine-pollution models (MariNeXt, Marine Debris Detector) — fused by confidence-and-voting logic, completed with classical CV post-processing, and fed into OpenDrift/OpenOil oil-slick drift simulation with live tidal and weather inputs.
Metrics, coverage, and customer context are confidential; the techniques are not — ask me about it.
FROM PROBLEM TO PRODUCTION.
The same loop drives my engineering and my pre-sales work. Five steps. No magic. Repeatable.
Discover the real problem.
I start where the customer is — not where the technology is. Enterprise discovery taught me to listen for the real problem underneath the stated one, map stakeholders, and find the constraints that actually matter.
Define the architecture.
Design against reality: the right model, the right integration pattern, the right deployment. Then explain the trade-offs in whatever language the room speaks — engineering or executive.
Build & validate.
Working software in days, not months. Whether it's a customer-facing demo, a POC, or a full production system — I write the code myself and get it in front of real users early.
Explain, demonstrate, ship.
A system nobody understands doesn't get bought — or adopted. I turn architectures into demos and narratives that survive both a technical evaluation panel and a boardroom.
Measure & improve.
Nothing matters until it works in the real world. I run evaluations, measure outcomes, and iterate on actual usage — the goal is production value, not a POC that lives in a folder.
References and the long-form PDF available on request: daniel@mangelsdorf.au