Build the software. Secure what it runs on. Keep IT running.
We build and maintain your software. We look after the IT and security it runs on, so a small firm does not need a team of its own. And when a new vulnerability lands, we read it against the vendor’s own advisory and write down what to do about it. Each of the three leaves a record you can check.
- AI-operated delivery
- Specialist AI roles do the routine work, so a small team can take on what would otherwise need a large one.
- Human accountability
- A Technical Architect owns what ships, and a Forward Deployed Engineer works with you and watches your running product.
- Evidence at every step
- The checks that ran, the decisions and why, and the board we ran it on are yours to keep and to check.
Our mission is to help small and mid-size businesses secure their IT assets, and to help startups afford fast, secure digital solutions for growth.
NamiQ LLC, Austin, Texas. Founded by Hung Trong Le.
Software built and maintained, with a check on every change
- Features delivered
- 182/184
- Defects per 1,000 lines
- 2.64
Our own product, as at 2 August 2026. The unflattering rows are on the results page.
Every new vulnerability read against the vendor's own page, and a decision written down
IT and security looked after, for firms with nobody to do it
Help-desk targets in business hours, shown against a 10-hour day. Targets, not guarantees.
A new vulnerability is a question about your business. Here is how we answer it.
Not “how bad is the score” but “do we run it, who can reach it, what would it cost us, and by when”. In the ten CVEs we worked through on 28 October 2026 the score was a poor guide: a 2.7 was on the exploited list, and four unauthenticated 9.8s were not.
- 1Check
- 2Analyse
- 3Triage
- 4Act
- 5Prove
- Act nowExploited and reachable: mitigate in hours, patch inside the deadline
- This weekSerious, or exploited with a narrow reach: within seven days
- ScheduledReal but not urgent: the next patch window
- TrackA precondition does not hold today: watch for change
- Not affectedClosed, with the evidence written down
- Find outUnknown is not a state: inventory the asset today
We build and maintain your software, secure what it runs on, and hand you the evidence for every change.
A delivery service, not a tool
Run by AI, owned by people
You keep everything, including the reasoning
If one of these is yours, we have done it before.
You have something to build
You inherited something nobody understands
Nobody owns your infrastructure
An AI tool built it, and now it breaks
Our own product first, measured in public.
Read this first. One product, with the scope, the deadline and the stack chosen by us, and no client, procurement or change requests in the way. It shows how the factory runs. It does not show that your project will go the same way.
- “Work completed” counts items closed, not items proven by a person finishing a journey on the running system.
- Defects per 1,000 lines counts closed defect tickets, live and archived, over the lines in the product’s own directories.
- These come from the live cybersecurity platform we build on this factory, 165,000 lines, as at 2 August 2026, recomputed from the work record rather than typed into a slide. The same board reports that 43% of work was later redone and that queueing time rose as parallelism increased.
Who we help, who runs this, and what we have done for them.
You are entitled to check us. Here is who runs this, what we are doing for clients as at 28 October 2026, and the work you can read without talking to anyone.
Two decades building software for global clients in the United States, Japan and beyond, then five years building an AI business in Vietnam, before founding NamiQ LLC in Austin, Texas.
Hung Trong Le on LinkedInAn established, growing architecture firm in Austin, Texas, with eight to ten people, and a client since September 2026.
- We are its IT department: laptop installation, Google Workspace administration and a help desk.
- The help desk has a 2-hour first-response and 4-hour resolution target in business hours, written into the desk's service policy. They are targets, not guarantees.
- Its shared drives, over a terabyte across ten drives, are copied every night to a disk on the firm's own network, and each run reports to a dashboard we watch.
As at 28 October 2026: the last nightly copy completed, with all 19 jobs ok and none failed.
DTLN, a paying client with a Socratic AI learning platform of about 147,000 lines, on a different technology stack from our own.
- We take on enhancement and maintenance of an existing system we did not write.
- Its material is the client's, so by default it is handled on local models only. Any exception is written down with who approved it and why, and every use is logged.
This is the engagement behind the factory's second codebase on our results page.
Bootcivil LLC, a paying client since October 2026.
- We manage its IT and security on the same service as our other IT client, supported through the same help desk.
- We watch its estate using OCO, our cyber knowledge graph.
Managed since October 2026.
What you can read without talking to us
Counts are read from the Insights page as it stands, not typed. Where a source could not be read or two sources disagreed, the note says so.
The pages most people need, by the job.
Build software
- Building software →For founders and engineering leads whose AI-assisted codebase has gone wrong
- Inherited systems →The author left and the documentation is out of date
- How we build →Four steps, six checks and the evidence kept for every change
- Delivery record →What the factory produced on a live product, unflattering readings included
Secure and run IT
Know what to do about a new CVE
- Today's mission →What to patch now, in deadline order. Refreshed after every data update
- CVE Triage Handbook →63 pages on checking and triaging a CVE, free with a work email
- AI SDLC Playbook →Six practices, and a 58-page full edition, free
- Lattice →A public cyber knowledge graph: an analyst chat, an MCP server and a host check
This is not our claim. It is what 2026 published.
more issues in pull requests containing AI-assisted code than in human-written code
arXiv preprint 2603.28592, March 2026maintenance costs by year two, for teams that do not actively manage the debt
Innovative Group, 2026Published figures, directional rather than audited. Three of the four come from the same 2026 analysis; the first is a separate preprint. Every one is linked so you can weigh it yourself.
Two ceilings nobody warns you about.
We wrote these in July 2026, from running a factory past 165,000 lines. The measurements since have pointed the same way.
Your reviewer is your throughput
It degrades exactly where you are going
More generation makes it worse, not better.
Buy more AI tooling
Hire senior engineers
Bring in an agency
Gates decide what ships. Not prompts, and not optimism.
Agents do the work; 6 checks decide whether it reaches your product. The last one asks whether a real person can finish the job on the running system — the failure tests cannot catch.
Understand
We read your problem — or your existing system — into a map of what it is, what depends on what, and where the gaps are. You get a plan you can challenge before anything is built.
Build
The work is split into small, tracked items with agreed acceptance criteria and assigned to the right specialist. Every change is isolated and reviewed.
Prove
Six checks run before anything reaches your product — ending with whether a real person can finish the job on the running system.
Report
Progress, quality and cost are recorded as the work happens. One live board, always current. Blocked work shows the day it blocks.
What we have learned running it.
Java4Shell, CVE-2026-114938: the OpenJDK Base64 flaw, which Java servers it can reach, and what to do this week
A memory-corruption bug in OpenJDK's Base64 decoder reaches Jira, Keycloak and Jenkins, but only on x86-64 CPUs with AVX-512 VBMI. How to tell, and what to do.
nginx2root: the five nginx flaws fixed on 18 November, which of them reach your servers, and what to do first
nginx 1.9.5 to 1.31.8 with HTTP/2 on is open to CVE-2026-123007, an unauthenticated memory corruption, unless a backport fixed it. Four more need set-ups.
Daily CVE highlight, 19 November 2026: 3 new critical CVEs worth reading
260 new records scored 8 or higher in the last 30 hours. 3 to read first, and the conditions each score leaves out.
Start with one piece of work.
Give us one piece of work and read access to one system. We map it and come back with what we found, what we would build and what it would take, at no charge and before any commitment. About a week from read access to write-up. One service or repository up to roughly 50,000 lines; a larger estate we scope first rather than promise a week for.
Talk to us