Sparli Aman · production agent engineer
I wire tool-using agents into CRMs, telephony, repos and regulated comms. I get called when one is doing something wrong.
7y client-side delivery live in US · CA · UK UTC+5 full EU overlap
sparli aman · pakistan → everywhere How my agents run in production
live loop · css only, zero js
01The pattern
Not caution: asymmetry. You cannot recall a bad message to a customer, or unspend an ad budget. I have made this call three times, in three unrelated domains, without anyone asking me to.
confirm_action, fully audited.
pending_actions row and wait. The send path does not exist rather than
being disabled.
prepared. submitted is a
human attestation, and nothing reaches it without one.
02Case studies
Client work under NDA, so these describe architecture and outcomes rather than source. The frames below are engineered for screen recordings; until those are exported, each shows the actual architecture, drawn and moving.
SEO SaaS, Canada · 2025
The prospect said “hello”. Speech-to-text, working on a noisy first packet, heard “no”, and a perfectly reasonable end-call rule did the rest. Two correct behaviours composing into a wrong one. The logs were structurally incapable of showing it: every line in them was correct. Found by listening to the recordings.
Croncore internal, daily production use · 2025
A prompt saying “always ask before sending” is a request, and its failure is silent. So the send capability does not exist: no tool emails a client, no tool posts to ClickUp. Proposed actions land in a table and wait for a human. The worst case stops being a wrong message to a customer and becomes a badly worded row that gets rejected.
Telephony, language services client · 2024
The client needed Mongolian speech in a running call flow. No commercial vendor offered it, and the telephony side was a live Asterisk/FreePBX install whose audio contract could not change: sample rate, encoding, timing, all fixed. I built and operated the serving stack; I did not train the models, and I will say which part was mine before you ask.
A self-hosted FastAPI service running two off-the-shelf speech engines behind one interface, with fallback between them, producing audio that dropped straight into the existing call flow.
03In flight
Listed separately from shipped work on purpose. Neither has results yet, and a page that blurs the two is not worth reading.
04Method
I moved a production voice agent between models because latency on live calls was too high, not because of a leaderboard. Same instinct: stress-testing to 200+ concurrent users before release, not after.
A system prompt is a request. A missing code path is a guarantee. When the failure is irreversible, I build the guarantee.
Dialling windows, DNC scrubbing, consent logging, CASL unsubscribe: built in-house, per jurisdiction, because the penalty lands on the client’s phone numbers.
I build, integrate, deploy and operate these systems. I do not train models, and where a project involved one I will tell you which part was mine.
05In production with
Twilio · Asterisk/FreePBX · Deepgram · ElevenLabs · Pipedrive · ClickUp · GitHub · Gmail · Amazon Ads API · Meta Ads · Google Places · Stripe. Each wired up and debugged in production, not read about.
06 · Colophon
The claims above are the kind that should be checkable, so the site itself is a small demonstration. Every number below is enforced by a script that fails the build when it regresses, including the animation above, which cost zero bytes of JavaScript.
Astro static · hand-written CSS · a content schema that refuses to build a case study without an evidence field · CI actions pinned to commit SHAs · deployed through the same gates it advertises. Read the source.