Services Capabilities Sectors About Let's Talk →
Fractional AI/ML Consulting
R&D
Production

Help your lab iterate 10–100x faster

I bring automation, real-time sensing, and machine learning to chemistry and materials R&D. Senior expertise on a fractional basis — without the full-time cost.

Your process knowledge, my engineering

You have the instruments, the domain knowledge, and the processes. I bring the engineering, coding, and ML to make your existing setup faster, smarter, and easier to keep compliant.

Let's Brainstorm → See How I Work
MSc Mathematics, Cambridge
PhD Theoretical Physics, Groningen
MIT — Bioprocess Data Analytics & ML
10+ years applied ML
Services
Three ways to work together
Start small and expand as needed. Most clients begin with an Assessment to map opportunities before committing to larger engagements.

Assessment

1–2 weeks

Map opportunities and get a prioritized roadmap for automation and ML in your R&D workflows.

  • Review of instruments, data, workflows
  • Opportunity ranking by impact
  • Recommended first project
  • Build vs. buy guidance

I spend time understanding your instruments, data flows, and bottlenecks — then map where automation and ML will have the biggest impact.

  • Walk your process, learn your pain points
  • Identify data that's collected but unused
  • Spot manual steps ripe for automation
  • Prioritized improvement roadmap
Popular

Development Sprint

12 weeks

Build something real. Weekly sessions plus async support to create a working prototype.

  • One focused project, start to finish
  • Weekly working sessions
  • Documentation and handoff
  • Your team learns alongside

Build one working system: a dashboard, ML model, data pipeline, or automation that solves a concrete production problem.

  • Get data out of instruments into usable systems
  • Automate reporting, trending, or analysis
  • Build ML models on your process data
  • Documentation and team handoff

Fractional Leadership

Ongoing partnership

Embedded technical leadership. 1–2 days per week as your AI/ML lead.

  • Architecture decisions
  • Team mentoring
  • Technical roadmap ownership
  • Investor/board communication

Ongoing engineering and data science capacity. 1–2 days per week — I become the person who makes your data actually work for you.

  • Build and maintain data infrastructure
  • Automate manual analysis workflows
  • Train your team on tools I build
  • Evaluate vendors, guide technical decisions
Capabilities
What I build for R&D
I don't just add ML to your workflow — I redesign the whole system for maximum data value and fast iteration cycles.

Real-Time Monitoring

UV/Vis, Raman, hyperspectral with millisecond feedback. Closed-loop control for synthesis and characterization.

ML Optimization

Bayesian optimization, active learning, experiment design. Reduce 1000 experiments to 50.

Hardware Integration

Microfluidics, flow chemistry, legacy instrument retrofits with IoT connectivity.

Data Infrastructure

Sensor-to-cloud pipelines, interactive dashboards, automated spectral analysis.

Physics-Informed ML

Custom models that encode domain knowledge. Work with limited data, extrapolate reliably.

Nanoparticle QC

Real-time size, shape, and quality prediction from spectroscopy. Replace slow TEM with fast ML.

What I build for production
Your team knows the process and the regulations. I bring the engineering to get data flowing, automate the tedious parts, and make compliance easier — not harder.

Get Data Out of Instruments

Connect legacy machines to modern systems. Export proprietary formats, stream to dashboards, make data accessible and searchable.

Automated Trending & Reporting

Replace manual Excel reports with live dashboards. Batch comparison, deviation alerts, historical trends — updated automatically.

ML on Your Process Data

Predict batch outcomes from early indicators. Find patterns humans miss. Models trained on your data, for your specific process.

Make Compliance Easier

Build data pipelines with audit trails, timestamps, and traceability baked in. Your quality team defines what — I build how.

Automate Manual Steps

If someone is copy-pasting data, manually checking spectra, or filling in spreadsheets — I can probably automate it.

Real-Time Monitoring

Spectroscopy, particle sizing, sensor data — streamed live to dashboards with alerts. Catch problems during the batch, not after.

Sectors
Where I work in R&D
Same core skills — real-time sensing, ML, automation — applied to different chemistries.

Nanoparticles & Materials

Synthesis optimization, characterization, size/shape control, novel morphologies.

Biotech & Pharma

Cell line development, formulation screening, crystallization monitoring.

Catalysis

Catalyst screening, operando spectroscopy, activity/selectivity optimization.

Energy Materials

Battery materials, perovskites, electrode characterization, bandgap mapping.

Industries I've worked in or am targeting
I learn your specific process quickly. The engineering patterns transfer across industries. Your domain experts guide the what; I build the how.

Biopharma

Batch monitoring, sensor integration, making compliance data infrastructure painless. MIT-trained in bioprocess analytics.

Specialty Chemicals

Reactor monitoring, yield tracking, nanoparticle QC. Get data out of instruments and into decisions.

Food & Beverage

Inline quality checks, composition monitoring, automating manual QC steps. Spectroscopy is spectroscopy.

Chemical Production

Process monitoring, sensor data pipelines, predictive models for yield and quality. Legacy instruments welcome.

About

I'm the technical accelerator your team is missing

I'm Sjoerd de Haan. Physics background, not chemistry — which is actually the point. I see instrumentation and data problems that domain experts work around every day because they don't have the engineering bandwidth to fix them.

My pattern: I come in, learn your process and instruments quickly, then build the automation, data pipelines, and ML models your team needs but can't build themselves. I've done this across medical diagnostics, computer vision, engineering data, and analytical chemistry. The domain changes; the approach doesn't.

MIT Professional Education — Bioprocess Data Analytics & ML

Trained by Prof. Richard D. Braatz (MIT, National Academy of Engineering) in process data analytics, sensor technologies for manufacturing, and ML methods for bioprocess operations.

I work on a fractional basis: typically 1–2 days per week with each client. You get senior technical expertise without the cost (or commitment) of a full-time hire. I build things, hand them over, and make sure your team can maintain them.

Sjoerd de Haan
LinkedIn GitHub Blog
Get started
Have a thorny R&D problem?
I enjoy thinking through hard problems with creative teams. Let's brainstorm how automation, real-time sensing, and ML could shift what's possible for you.
Drowning in data but not learning from it?
Tell me what instruments you have, what's annoying about your current data workflow, and what you wish worked better. I'll tell you honestly what I can help with.
Pick My Brain →

or email sjoerd@gogrowai.com · +31 6 1226 1747