Streaming & real-time
You’re seeing
Batch pipelines hitting latency limits. Event volume outgrowing ingestion. Kafka adopted, and painful.
We deliver
Event ingestion on Kafka with managed schemas, monitored consumer lag, and replay you can trust.
Senior data engineering · Embedded or scoped · For Dutch product teams
Your data roadmap shouldn’t wait on a hire you can’t make.
FulmenFlux puts senior data and backend engineers into your team for streaming, AI/RAG infrastructure, pipeline reliability and migrations. Start with a fixed-price Discovery Sprint, and keep the findings whether or not you continue with us.
01 — The capacity problem
The work is clear. The people to do it aren’t there.
A
The senior data role has been open for months.
Few candidates reach the final round. The recruiter invoice arrived anyway.
B
One engineer owns every pipeline.
If they leave next month, nobody else knows why the nightly load breaks or how to fix it.
C
This quarter’s roadmap slipped on data work.
Product features wait on ingestion, backfills and jobs that fail at 3 a.m.
D
Customers spot broken data before you do.
Support hears first, someone digs through logs, and the fix takes days.
A failed hire costs the recruiter fee, months of salary, and the quarter you spent waiting. Put that number next to our rate card.
02 — Why now
Teams call us when there’s a date on the calendar.
No deadline in the next 90 days? Then we’ll probably tell you to wait, and check back when there is one.
- 01A funding round with a hiring plan recruiting can’t keep pace with.
- 02A compliance deadline: an audit, the EU AI Act, or a customer’s security questionnaire.
- 03A customer-facing data product with a launch date.
- 04An acquisition, and another company’s data to integrate.
- 05A directive to cut cloud data spend or headcount.
03 — What we do
Four kinds of data work. Find yours by the symptom.
AI & RAG infrastructure
You’re seeing
An AI feature mandate from the board. A prototype that works in a notebook but not in production. Indexes going stale.
We deliver
Ingestion and embedding pipelines, index refresh on change, access control and evaluation before launch.
Reliability & governance
You’re seeing
Pipeline failures reaching customers. No data contracts or lineage. Dashboards nobody trusts.
We deliver
Data contracts, tests at every boundary, lineage, and alerts that reach a named owner first.
Migration & cost
You’re seeing
A legacy warehouse. An Airflow move nobody has time for. A cloud data bill that keeps climbing.
We deliver
Phased migration with parallel runs, a tested cutover plan, and cost per pipeline you can see.
Where our depth is, and where it ends.
Tell us your stack on the first call. If we can’t staff it credibly, we’ll say so then, before anyone spends a euro.
Deep, in productionGo · Kafka · PostgreSQL
Also in the stackPython · Spark · Airflow · dbt
LakehouseSnowflake · Databricks, chosen per engagement
DomainsFintech · healthtech · developer tools · SaaS
04 — How we engage
Start small. Commit when you’ve seen the work.
Step one
Lowest risk
Discovery Sprint
2–3 weeks · fixed price €7,500
We review your pipelines, stack and recent incidents, then hand over a written findings report and plan. It’s yours to keep, whoever ends up building it.
Best when you don’t know us yet, or the budget isn’t allocated.
One outcome
Scoped project
Fixed price or time & materials
A migration, a streaming pipeline, a RAG service in production. Fixed price when the scope is clear; time and materials when it isn’t. You choose, and we explain the trade-off.
Best for a defined deliverable with a deadline.
Ongoing capacity
Dedicated pod
3-month minimum, then monthly
A tech lead and 2–4 senior engineers join your standups, your repo and your on-call, and answer to your product owner. The same people for the whole engagement, never a rotating bench. Why three months: onboarding takes a few weeks, and a shorter stint pays for onboarding and gets little after it.
Three months feels long for a new vendor? Do the Sprint first.
Hours
Your full working day. Moldova is one hour ahead of the Netherlands.
Daily contact
One named lead on our side. On yours, someone with a few hours a week to onboard us in the first two weeks.
Access
Your VPN, your device policy, least-privilege repo access. Production access only when the work needs it, and you approve it.
Handover
Docs and runbooks written as we go. Your team owns the code from the first commit.
05 — Rate card
What it costs, before you ask.
The prices are published and fixed. If the budget is tight, we change the scope, the team or the contract length, never the rate.
| Discovery Sprint2–3 weeks, fixed fee | €7,500 |
|---|---|
| Starter pod1 tech lead + 2 engineers, per month | €22,800 |
| Growth pod1 tech lead + 3 engineers, per month | €29,800 |
| Scale pod1 tech lead + 4 engineers + QA, per month | €42,000 |
Invoiced in EUR, net 30. Annual indexation of 4–6%, written into the contract on day one, so nothing new appears in month 13. The one discount is 5% for a 12-month prepay, the same for every client.
What you’re probably comparing us with
Hiring in-house
Cheaper over years, once the hire lands. If you can hire in weeks, do that. We cover the months until then and hand over cleanly.
Freelancers
Often a lower day rate. When they leave, nobody replaces them; with us, a 21-day replacement guarantee is in the contract.
Polish or Romanian firms
A similar model. Ask them for pod prices in writing and for the names of the people who would be on your team.
“Is it cheap because it’s junior?”
The rate reflects Moldovan salaries, not seniority. You interview every engineer before they start, and the Sprint lets you judge the work first.
06 — Delivery risk
What happens when something goes wrong.
You’ve had a vendor go badly before. Here’s what’s in our contract for the usual ways it does.
An engineer leaves mid-project
A 21-day replacement guarantee is written into the contract. Handover overlap is at our cost, so you don’t pay for ramp-up.
Communication isn’t working
Our engineers work in English every day, and you interview each one. Say “this isn’t working” in week two; we’d rather hear it then than in month three.
Scope creeps
Written scope. Change requests are priced before work starts. One accountable lead on our side.
The engagement ends
The handover plan is in the contract: docs, runbooks, recorded walkthroughs, and a transition period.
07 — Trust, legal & GDPR
A young vendor from Moldova. Fair questions, plain answers.
Is Moldova stable and EU-aligned?
Moldova is an EU candidate country and opened accession negotiations in 2024. Our engineers work remotely, and your code and docs live in your own repo, so the work doesn’t depend on one location.
How does GDPR work with a non-EU vendor?
Moldova has no EU adequacy decision, so we sign a DPA with Standard Contractual Clauses and share our subprocessor list. Your data stays in your EU cloud; we work inside it.
Who owns the work?
You do. All work product and IP is assigned to you on payment.
ISO 27001 or SOC 2?
We don’t hold either certification yet. We send our security controls list and answer your questionnaire in writing.
What if you disappear?
Code lives in your repo and docs in your wiki from day one. Nothing you run depends on us staying in business.
Do you have Dutch references?
Not yet. What we can show is ABAXUS, the real-time billing engine our founder built and still runs in production. The Sprint is how you check us without a long commitment.
08 — For everyone who signs off
Several people will weigh in. Each gets what they need.
CTO / Head of Data
Sprint findings, engineer CVs, a technical interview with each engineer.
CFO / Finance
Itemized rate card, a failed-hire comparison, and options for headcount or project budget.
Procurement & Legal
MSA, DPA with SCCs, subprocessor list, IP clause, insurance certificate.
Security
Controls list, access model, your questionnaire completed.
Your engineers
They review our PRs, pair with us, and own the result. We add capacity; the system stays theirs.
09 — When we’re the wrong choice
We’d rather say no on the first call than in month two.
- Nothing needs to change in the next 90 days.
- There’s no budget path yet. We’ll talk again when there is.
- Nobody on your side can own the work technically.
- Your stack is one we can’t staff well.
- The deal only works with a large discount.
10 — What people ask next
The rest of the objections.
We prefer to hire and build in-house.
Good. We’ll cover the gap until your hire starts, then hand over to them.
Won’t AI tools cover this soon?
They help write pipeline code. Someone still has to own the schemas, the 3 a.m. failures and the cloud bill. That’s the work we take on.
We have no bandwidth to onboard anyone.
The Sprint needs a few hours of your team’s time in total: an architecture walkthrough and access to your pipeline docs. We’ll tell you the exact ask before you sign.
We already have a vendor.
Then run a Sprint on the part they don’t cover, and compare the results.
Can we start with less than three months?
Yes: the Discovery Sprint or a scoped project. The minimum applies only to dedicated pods.
Not now. Maybe next quarter.
Fine. Tell us the date that matters and we’ll check in a month before it.
What changes in the next 90 days if this isn’t solved?
Tell us in a couple of lines. Within one business day we reply with whether we can help and who would do the work.
Prefer to talk? Book 30 minutes directly