KK
Founder & engineer
Platform, voice pipeline, public face for partners.
About
We're building voice AI for the calls people actually take.
Calleague is a small, opinionated team that ships real-time voice infrastructure for regulated, high-volume work — where latency, languages, and trust matter as much as the model.
Our story
Calleague started inside Arpanet — a Turkish enterprise software house — when our customers kept asking the same thing: "Can your AI answer the phone?" The off-the-shelf options were either too slow, too cloud-locked, or too generic for what real call centers, sales teams, and clinics actually needed.
So we built our own pipeline: STT in 32 languages, low-latency TTS, an LLM router that picks the right model per call, and integrations that read and write to the systems people already pay for. Self-hosted from day one, because that's what regulated buyers asked for.
We ship in small steps and validate against real audio. Every release goes out with a public changelog, an honest WER table, and a list of what we haven't solved yet. That's the bar.
What we care about
These show up in every release note, every roadmap call, and every refusal to ship something half-baked.
Practical over flashy
We build what teams actually deploy. If a feature doesn't survive a real pilot, it doesn't ship.
Honest by default
Public changelog, honest benchmarks, clear gaps. No "AI magic" claims we can't back with numbers.
Quiet ambition
We're not the loudest in the category — we just want every call answered, in every language, with sub-second response.
The team
Small, focused, multi-disciplinary. Engineers, voice researchers, and people who've actually run call centers.
VR
Voice research lead
STT and TTS quality, multilingual coverage, latency.
PL
Platform lead
Self-host, on-prem, observability, SDKs.
DS
Product design
Studio UX, agent authoring, dashboards.
GT
Go-to-market
Pilots, partnerships, customer success across pilots.
SE
Security & compliance
SOC 2, GDPR, KVKK, HIPAA-readiness, BYOK.