SOSLAN ALDATOVVOICE AI & VOIP ARCHITECT

CASE STUDIES  /  PRODUCTION SYSTEMS

What I built, how it works, and what changed.

Detailed, anonymized breakdowns of the voice platforms I have designed, rebuilt and operated — including the architecture, the difficult engineering decisions and the production outcome.

~100K CALLS / 8HMULTI-TENANT VOICE SAAS

~30K CALLS / DAYAI-FIRST FREIGHT

~8M CALLS / DAYGLOBAL CLOUD PBX

~200K DEVICESCONNECTED CARE

SELECTED CASE STUDIES

Production systems, not generic portfolio pieces.

01 / VOICE AI FOR FREIGHT

~30K calls/day. Every inbound call is answered first by AI, then verified, qualified and routed. Drachtio · FreeSWITCH · Deepgram · Vapi · EKS

~30K calls/day; every inbound call first answered, verified and qualified by AI.

READ CASE STUDY →

02 / GLOBAL CLOUD PBX

Modernized a voice core carrying ~8M calls/day: AGI → ARI, services, Kubernetes, monitoring and Janus video. OpenSIPS · Asterisk · ARI · K8s

~8M calls/day; AGI → ARI services, Kubernetes, monitoring and Janus video.

READ CASE STUDY →

03 / MULTI-TENANT VOICE SAAS

Built a shared inbound/outbound platform from zero: ~100K calls per 8-hour day, ~5K peak sessions and instant tenant provisioning.

~100K calls per 8-hour day on shared infrastructure built from zero.

READ CASE STUDY →

04 / CONNECTED CARE COMMUNICATIONS

Shared audio/video infrastructure for ~30 care facilities and ~200K devices. MQTT · SIP · FreeSWITCH · RTPengine · Janus

Audio/video for ~30 facilities and ~200K wearable and stationary devices.

READ CASE STUDY →

OPEN-SOURCE PROOF

Small, inspectable pieces of the same systems thinking.

HAVE A SYSTEM LIKE THIS?

Bring me the architecture, the failure mode or the production target.

I can audit an existing voice stack, rescue a fragile deployment or own the path from SIP edge to real-time AI.

Discuss your project