Python Developer — AI Voice Agent + Windows Screen Automation (Healthcare, LLM, Local AI) - Contract to Hire
We are building ReadyRep.ai — a Windows desktop application (.exe) that acts as an AI front-desk receptionist for medical clinics. It answers phone calls using a local AI voice agent and simultaneously navigates an EHR system on screen (moving the cursor, clicking, entering data) — exactly like a human receptionist would.
What makes this unique: instead of a developer programming the steps, the clinic user performs each task once on their own computer while the app silently watches and records. The AI learns from that recording and repeats the task independently on future calls. No coding required from the end user.
We need a developer who can build the following 4 components in 4 weeks:
1. A clean Windows dashboard (Tauri + React) with a task library, a “Start Shadowing” recording mode, and a live status bar showing what the AI is doing step by step
2. A shadow recorder that captures mouse clicks by element name (not pixel coordinates) using the Windows UI Automation API via pywinauto — plus microphone audio transcribed locally via Whisper
3. An AI replay engine that loads the recorded task, substitutes live caller data using Jinja2 templates, and executes the steps autonomously using pywinauto — running in a parallel thread so it never blocks the voice conversation
4. A voice agent using Twilio (inbound calls), whisper.cpp for real-time speech-to-text, LLaMA 3 via Ollama for intent classification, and Kokoro TTS for natural voice output — with a filler phrase system that fills EHR loading delays so the caller never hears dead silence
The POC requires one working demo task: New Patient Enrollment — live phone call, fields fill in OpenEMR on screen in real-time, confirmation email sent via Resend API before the call ends.
Required skills: Python, pywinauto / Windows UI Automation, LangChain, Ollama / LLaMA, Whisper STT, Twilio Voice SDK, Tauri or Electron, asyncio (parallel processing), SQLite
Nice to have: Kokoro TTS experience, OpenEMR familiarity, HIPAA awareness
Budget: upto $5,000 fixed price, paid in two milestones (Week 2 checkpoint + Week 4 final delivery)
Deliverables: Working .exe installer, all source code in a private Git repo, and a live demonstration of the demo task. A full technical specification document will be shared with shortlisted candidates.
To apply, please answer these three questions in your proposal:
1. Have you built anything using pywinauto or Windows UI Automation before? Briefly describe it.
2. Have you worked with a local LLM (Ollama, LLaMA, Mistral) in a production or client project?
3. What is your realistic timeline to deliver a working voice + screen automation prototype?
Proposals without answers to these questions will not be reviewed.
Apply Now
Apply Now
What makes this unique: instead of a developer programming the steps, the clinic user performs each task once on their own computer while the app silently watches and records. The AI learns from that recording and repeats the task independently on future calls. No coding required from the end user.
We need a developer who can build the following 4 components in 4 weeks:
1. A clean Windows dashboard (Tauri + React) with a task library, a “Start Shadowing” recording mode, and a live status bar showing what the AI is doing step by step
2. A shadow recorder that captures mouse clicks by element name (not pixel coordinates) using the Windows UI Automation API via pywinauto — plus microphone audio transcribed locally via Whisper
3. An AI replay engine that loads the recorded task, substitutes live caller data using Jinja2 templates, and executes the steps autonomously using pywinauto — running in a parallel thread so it never blocks the voice conversation
4. A voice agent using Twilio (inbound calls), whisper.cpp for real-time speech-to-text, LLaMA 3 via Ollama for intent classification, and Kokoro TTS for natural voice output — with a filler phrase system that fills EHR loading delays so the caller never hears dead silence
The POC requires one working demo task: New Patient Enrollment — live phone call, fields fill in OpenEMR on screen in real-time, confirmation email sent via Resend API before the call ends.
Required skills: Python, pywinauto / Windows UI Automation, LangChain, Ollama / LLaMA, Whisper STT, Twilio Voice SDK, Tauri or Electron, asyncio (parallel processing), SQLite
Nice to have: Kokoro TTS experience, OpenEMR familiarity, HIPAA awareness
Budget: upto $5,000 fixed price, paid in two milestones (Week 2 checkpoint + Week 4 final delivery)
Deliverables: Working .exe installer, all source code in a private Git repo, and a live demonstration of the demo task. A full technical specification document will be shared with shortlisted candidates.
To apply, please answer these three questions in your proposal:
1. Have you built anything using pywinauto or Windows UI Automation before? Briefly describe it.
2. Have you worked with a local LLM (Ollama, LLaMA, Mistral) in a production or client project?
3. What is your realistic timeline to deliver a working voice + screen automation prototype?
Proposals without answers to these questions will not be reviewed.
Apply Now
Apply Now