Shipped · Mac App Store
ColumnLens
SQL, charts and 3D visualisation for the files your other tools gave up on. CSV, JSONL, Parquet, Excel and SQLite — resolved in seconds, entirely on your own machine.
- Open 5.2 GB
- 2.8s
- Resident
- 683MB
Independent software engineer · Hannover
The software comes to your data.
Not the other way around. I build native tools and governed data access for people whose files are too large, whose fleets are too distributed, and whose data is too sensitive to hand to somebody else's cloud.
When I'm going through reams of logs I prefer a delightful UX over the CLI and this tool filled that gap for me. I looked far and wide for a fast and beautiful app before having the pleasure of coming across this one.
Products
Two instruments,
one conviction.
Everything here resolves your data where your data already is. No ingest pipeline, no multi-tenant pool, no telemetry.
-
-
In production
IoT Data Flow
The managed data box for machine fleets. One hardened binary per customer, physically isolated, never multi-tenant. Leadership asks in Claude, finance refreshes Excel, engineers query SQL.
- Machines
- 3,900
- Tenants
- 40+
Engineering work
Two things,
taken selectively.
I work alone on purpose. A short list of things done properly rather than a long list subcontracted.
-
Fleet data onboarding
Getting your machine data to answer questions: provisioning, connecting your databases read-only, curating the datasets that matter, and building the tools your team will actually ask through. Ends with your box live and your people using it.
- One week · fixed scope
- Continues as a monthly subscription
-
Native performance work
When the tool everyone else uses gives up. Desktop applications and data paths measured in milliseconds and megabytes, not in "should be fast enough".
- Scoped per project
- Starts with a measured baseline, ends with a measured result
Research
Working notes,
not thought leadership.
What I'm measuring, breaking and rebuilding — written down while it's still fresh enough to be useful.
-
Local-first data
Why a laptop with DuckDB beats a cluster for most analytical work, and the specific point where that stops being true.
-
Agents on real data
Giving an LLM access to a production database without giving it the database: curation, scoped tools, audit, and what actually goes wrong.
-
Native performance
Measurements from shipping C++ desktop software in — SIMD parsing, immediate-mode UI, and where the milliseconds really go.
Who you're hiring
One engineer,
named, in Hannover.
I've been an independent software engineer since . Before that, years of building IoT systems for companies that needed sensors, databases and people to agree with each other.
When you hand read-only credentials to a dedicated box, the honest question is who exactly can touch your data. The answer is me — and every query is logged, so you never have to take that on faith.
I believe professional tools should be fast, private and beautiful. Not two out of three.
Tell me what's missing
I build what
people ask for.
One list for everything I make. Release notes, measurements, and the occasional question about what to build next. Roughly monthly — reply to any of it. A good share of what ships started as somebody's complaint.
No tracking pixels · one-click unsubscribe · never shared
Get started
Bring a problem
worth measuring.
Tell me what's slow, what's stuck, or what your leadership keeps asking that nobody can answer. I'll tell you honestly whether it's a week of work, a quarter, or something you shouldn't pay anyone to do.
Marc Hildmann · Gehrden, Hannover Independent software engineer VAT ID DE 244263257