Performance audit & fix sprint · [2] slots left in [November]
Find what's slow. Get the code to fix it.
For teams dealing with slow API requests, database saturation or report timeouts. I measure five agreed pages or endpoints, implement fixes and deliver the benchmark results.
€2,500 fixed · 3 working days after setup
Two problems I've worked on
- production · multi-tenant SaaS
Issue-list API response time
−46%
median response time
I moved a repeated COUNT query to a separate cached endpoint. Node.js, Strapi v4 and MySQL.
Compared with the same weekdays a week earlier. p95 improved 12%. API timing, not full page-load time.
- controlled test · PDF generation
A batch of ten report documents
−89%
batch time · 9 of 10 completed, up from 4
I limited Chromium concurrency, resized report images and closed pages after failed renders.
Same ten documents under production-like CPU and memory limits. One still failed. Not a production success rate.
A bounded sprint. Reviewable changes.
I write the fixes; your team reviews, merges and deploys them through your usual process. Strongest fit: Node.js, Strapi and SQL-backed SaaS.
€2,500
Fixed scope · Three working days after setup
- Five agreed pages or endpoints measured
- Implementation within the three-day scope
- Pull requests with verification notes
- Before-and-after report, including remaining issues
At least one agreed target 30% faster, or you don't pay.
We agree the five targets, the metric and the benchmark setup before we start.
Setup first. Then three working days.
- before day 1
Agree the scope and get access
We choose the targets; you provide repository access, logs and a runnable environment with representative data.
- day 1
Measure and investigate
I establish the baseline and trace the slow work through the application, database and infrastructure.
- day 2
Implement the fixes
I make the changes in pull requests and verify them against the agreed workload.
- day 3
Rerun and hand over
You get the benchmark comparison, the pull requests and a report of what remains.
You work directly with Merlijn.
I'm a freelance backend engineer in the Netherlands who has worked on production systems since 2021, and I lead a development team today. My write-ups show the numbers before and after each fix, so you can see how I work before we talk.
Practical questions
Which stacks do you work with?
Node.js, Strapi and SQL are my strongest fit. For another stack, describe the application and symptoms on the call or by email. We establish whether a three-day sprint is realistic before booking.
Do you need production access?
I need repository access and a runnable environment with representative data. Read access to production monitoring and logs helps establish the problem. We agree how to share data during setup. Your team retains control of production changes.
How is the 30% measured?
Before the baseline run we record a timing metric for each target: median or p95 response time for an API, a named browser metric for a page. Before and after use the same dataset, resources, concurrency and cache state. If no target gets 30% faster on its metric, you don't pay. The report shows all five targets, including any regressions.
What if the problem needs more than three days?
The sprint delivers the fixes that fit the agreed scope and records the unresolved issues. Any further implementation needs a separate agreement.
What happens after the sprint?
You keep the pull requests and report. Your team reviews and deploys when ready. Further help is optional and separately scoped. The written quote confirms VAT treatment and payment terms before you commit.
Tell me where your users wait.
Bring the slowest endpoint, a trace or just the symptoms. We'll use a free 30-minute call to decide whether the sprint fits.
— Merlijn