00About
The name is an argument.
Gati is a Sanskrit word for motion , specifically motion with a direction, the word you would use for the movement of a planet rather than the fidgeting of a person. Q is the quotient. A measure. Something you can put a number against and compare next quarter.
01Origin
Where GatiQ came from
Most of the delay in shipping software has nothing to do with writing it. It sits in the gap between deciding something and being able to release it — a manual test queue, an approval board that meets fortnightly, a staging environment that has drifted so far from production that passing tests prove nothing, a deployment runbook long enough to need a checklist. None of that is a hard engineering problem. All of it is expensive, and almost nobody measures it.
That gap is the thing GatiQ was set up to close. Not by asking people to work faster, which never survives contact with a release weekend, but by shortening the loop: smaller changes, a pipeline nobody is afraid of, interfaces written down before code is written against them, and instrumentation from the first sprint so the improvement is visible rather than asserted.
Hence the name. Gati is the motion — with a direction, not just a speed. Q is the quotient: the measure that turns a claim about velocity into a number you can hold us to. A consultancy that will not measure its own delivery is asking you to take its word for the only thing that matters.

Pronunciation
GUH-tee-cue. Two syllables and a letter.
02Approach
We are consultants who stay for the operating phase
A consultancy that leaves at launch never learns what it built. The bill arrives, the team disperses, and the client discovers in month four that the retry logic was wrong under load. We structure engagements so that the people who made the architectural calls are still around when production tests them — usually for at least one quarter after go-live.
It changes the decisions. An engineer who knows they will be on the pager writes different code from one who knows they will be on another account. That single incentive does more for quality than any process document we could hand you.
- Small teamsThree to six people, one lead, no matrix.
- Written decisionsADRs in your repository, not our wiki.
- Your toolsWe work in your board, your CI, your review process.
- Exit planned earlyHandover is designed in week one, not week forty.
Vision and mission
- 01Vision
A software industry where speed is a property of the system rather than a debt taken out against it — where teams ship weekly because their tooling, tests and documentation make weekly safe.
- 02Mission
We design, build and operate software systems for other companies, then hand them over documented and tested, so the client's own engineers can keep moving after we leave.
Both statements are short on purpose. If a mission needs a paragraph of qualifiers, nobody in the company can act on it on a Tuesday afternoon.
03Values
Six things we argue about internally
Not aspirations on a wall. A value is only real once it has cost something to hold — a deal declined, a date reset, an uncomfortable conversation had early instead of late.
- 01
Estimate honestly, then hold the line
A date we invented to win a deal is a date we will miss in front of your board. We estimate in ranges, show the assumptions behind them, and tell you the week the range moves — not the week before launch.
- 02
Write the boring parts first
Logging, migrations, backups, access control. They never demo well and they decide whether the system survives its second year. We build them into the first increment.
- 03
Leave the code better documented than we found it
Every engagement ends with architecture decision records, runbooks and a diagram that matches production. If your team can't operate it without us, we haven't finished.
- 04
Measure before optimising
Profiles, traces and a baseline come before anyone rewrites a service. Most performance problems turn out to be one query, one missing index, or one chatty API call.
- 05
Say the uncomfortable thing early
If the requested feature will not solve the stated problem, we say so in week one and propose the alternative. Quiet agreement is expensive for everyone.
- 06
Momentum is engineered
Fast teams are not teams that hurry. They are teams with short feedback loops, small changes, and a deployment pipeline nobody is afraid of. We build those conditions deliberately.
04Structure
How the company is organised
Four practices, not four departments. Every engineer belongs to one practice for craft — code review standards, hiring, internal talks — and to one client team for delivery. The practice lead does not run the project; the delivery lead does. Keeping those two lines separate is why we can staff a team with a data engineer for six weeks without anyone renegotiating a reporting structure.
There is no bench in the traditional sense. Engineers between engagements work on internal tooling or on the open-source libraries we maintain, and that time is budgeted openly rather than hidden inside client rates.
We are a distributed team. Meetings are scheduled inside the client’s working day rather than ours, and we travel for the parts that genuinely need a room — discovery workshops, cutover weekends, incident reviews.

- L1
Platform & operations
Everything runs here
Infrastructure, delivery pipelines, observability and the on-call rota. The layer every other practice deploys onto.
- L2
Product engineering
Owns the domain model
Application and interface work. Holds the domain model and the data contracts the rest of the system agrees to.
- L3
Data & machine learning
Reads and writes L2
Pipelines, warehouse, metrics and models. Sits downstream of product engineering and is only as good as the contracts above it.
- L4
Security & compliance
Reviews every layer
Threat modelling, testing and audit evidence. Reviews the other three rather than sitting beside them.
Practices are layers, not silos — one engagement usually touches three of them.