Company. 2024 to 2026. Co-founder and software architect.
SmartCoderLabs
An AI-assisted software engineering ecosystem: code generation and refactoring, automated testing, and delivery and quality controls wired into the same pipeline.
The problem
Agentic development produces more code than a team can review by reading it. The bottleneck moves from writing to trusting, and no amount of model quality removes that.
So the ecosystem treats model output the way it treats any other contributor: it only reaches main through the same gates. Generation and refactoring sit on one side, the pipeline that measures them sits on the other.
Model output becomes a measurable signal
SonarQube, Sentry and OpenTelemetry turn generated code into the same quality numbers the team already reads for hand-written code.
The pipeline is the reviewer of record
Nothing merges on the strength of having been generated. The gate does not know or care who wrote the diff.
Decisions and the price I paid for them
Agentic development ships behind quality gates
CI/CD, SonarQube, Sentry and OpenTelemetry turn model output into measurable software-quality signals instead of unreviewed volume.
Trade-off accepted
Gates slow a single change down. Without them, AI-generated volume is exactly what nobody can review.
Instrumentation is part of generation, not a follow-up
If telemetry is added after the fact it never gets added, and the first production question has no answer.
Trade-off accepted
Every generated service carries integration cost on day one, before anyone knows whether it will survive.
Inside the build
-
Fig. 02 Quality gate blocking a pull request on coverage and duplication.
-
Fig. 03 Telemetry: the same traces for generated and hand-written services.