Predictive Dialer & AMD Engineer- Ukraine

AI Case Leads connects people hurt in accidents with personal injury law firms across the US. We’re replacing our hosted dialer with one we own, built on FreeSWITCH and Python.

You’ll own answering-machine detection (AMD) and call-screening handling, working alongside the engineer building our dialing engine. The goal is simple: stop losing live conversations when a phone’s screening assistant answers first.

What you’ll build

• Detect live people, voicemail, call screeners, and silence from the first seconds of call audio.

• When a screener answers, play a truthful message with our name and reason for calling, then connect a live rep when the person picks up. Route voicemail to the right message drop.

• Work with our dialing engineer on pacing, abandoned-call controls, SIP/media handling, and CRM reporting.

• Build a labeled set of real call audio and test every detection change against it. Track accuracy by result, phone type, and campaign, with particular attention to live people incorrectly classified as machines.

Required experience

• Hands-on work with iPhone call screening, including test calls to real iPhones.

• Personally built or deeply extended a production predictive dialer, including pacing logic. Vendor API integration, administration, or agent-interface work alone is not enough.

• Improved and measured AMD in production, beyond enabling a vendor setting.

• At least 2 years of production experience with FreeSWITCH, Asterisk, Kamailio, or OpenSIPS.

• Ability to read SIP traces and troubleshoot RTP, audio, and carrier issues.

• Experience using speech-to-text, audio features, or machine learning to classify real call audio.

• Ability to ship production backend services. Our stack is Python/FastAPI, PostgreSQL, Redis, RabbitMQ, and React/TypeScript.

Nice to have

• Experience at a dialer or contact-center software company.

• FreeSWITCH module development in C or Lua, including mod_avmd.

• Low-latency speech-to-text and English/Spanish voicemail drops.

• Familiarity with TCPA, DNC scrubbing, STIR/SHAKEN, and caller ID reputation.

Role details

• Remote, full-time contract on a 12+ month roadmap.

• At least 4 hours of overlap with US Central time.

• Payment biweekly or monthly in arrears via Wise, Payoneer, or wire.

• NDA and full IP assignment required. Code and accounts stay in our name; lead data stays in our systems.

How to apply

• Complete the application below with your résumé, LinkedIn profile, and specific examples of your own work. We review the written answers, then invite strong applicants to a camera-on walkthrough of their telephony work and code. The final step is a short paid work sample on test call audio.