Responsibilities
- Build and maintain agent-runtime components for multi-turn financial conversations, including payment negotiations, compliance guardrails, and objection handling.
- Implement context management for consumer state, conversation history, business rules, and branching dialogues.
- Develop conversational reasoning primitives that balance natural dialogue with verifiable decisions.
- Work on a real-time voice pipeline targeting sub-one-second latency across transcription, reasoning, and synthesis.
- Tune voice activity detection and turn-taking logic, evaluate and integrate speech models, and debug streaming audio and WebRTC issues in production.
- Write prompts, build orchestration flows, contribute to model-routing logic, and improve the agent evaluation framework.
- Own features and components from day one and ship production functionality used in live consumer conversations.
Requirements
- 2–5 years of hands-on engineering experience with exposure to building or working with ML or AI systems in production.
- Strong Python fundamentals and the ability to write maintainable code and debug production issues.
- Experience with LLMs, including prompt engineering, API integrations, or simple pipelines and agents.
- Strong ownership, learning, problem-solving, and execution skills in a fast-moving environment.
- TypeScript familiarity is a plus.
- Voice AI, speech recognition, speech synthesis, telephony, Twilio, LiveKit, streaming protocols, WebRTC, low-latency system design, fintech, lending, collections, or agentic-system experience is preferred.
Benefits
- Real ownership from day one, with the opportunity to own components and ship features used in live consumer conversations.
- Direct mentorship from senior engineers and the AI Lead.
- Exposure to frontier applied voice-AI problems and high-impact work affecting millions of financial conversations.
- Fast-paced, intellectually stimulating environment with opportunities for significant professional growth.
Tech Stack
Categories
About Prodigal
Prodigal maximizes payments for lenders and debt collectors by building dynamic strategies and motivating consumers with highly engaging, personalized treatments. Our advanced genAI has been trained on over 400 million consumer finance conversations, delivering unmatched industry expertise so you can drive record recovery rates. Experience the power of intelligent debt resolution with Prodigal’s AI that pays. Prodigal is headquartered in Mountain View, California, and our global team is on a mission to build the intelligence layer that powers consumer finance. With the backing of domain experts, technology leaders, and top investors, including Accel, Menlo Ventures, and Y-Combinator, Prodigal is poised to become the next iconic vertical SaaS company.