About the Role

You will be responsible for ensuring the performance, scalability, and continuous improvement of our AI systems. This includes end-to-end AI pipeline orchestration (STT - LLM - TTS), implementing and monitoring prompt flows, fallback strategies, and recovery logic. You will also focus on measuring and enhancing conversational quality, including latency, hallucination, success rate, and barge-in. This role involves building test sets, conducting A/B experiments, and preventing regressions before they impact users. Collaboration with the Tech Lead and Voice UX designer is crucial to align model behavior with real-world conversations.

What You Bring

• Solid Python programming skills.
• Experience integrating with LLM APIs (e.g., OpenAI, Anthropic).
• Proven experience shipping or maintaining LLM-powered products in production.
• Familiarity with STT/TTS tools such as Whisper, ElevenLabs, Coqui, Google, and Azure.
• Experience in prompt engineering, context control, session state management, and fallback orchestration.
• Ability to log, trace, and debug model behavior akin to backend logic.

Bonus Points

• Experience with Langfuse, TruLens, Helix, Ragas, or building custom evaluation tooling.
• Experience with latency-sensitive systems, including real-time or streaming pipelines.
• Experience setting up CI for prompt evaluation or contributing to LLMOps pipelines.
• Experience managing golden sets, live traffic shadowing, or prompt regression testing.

What You'll Get

• Significant ownership over AI behavior and quality across the product.
• A collaborative team environment that values pragmatism, iteration, and clean architecture.
• Opportunity to influence the voice, cognition, and adaptability of AI agents.
• The chance to shape an AI system that directly interacts with users and is evaluated on every interaction.

What This Role Is Not

• It is not a role focused on basic prompt testing or aesthetic quality assurance.
• It does not involve training foundation models or fine-tuning large language models from scratch.
• It is not a research-oriented position.

This position is ideal for someone who is motivated to deploy robust LLM-powered voice agents that perform reliably under real-world conditions and demonstrate continuous improvement.

How to Apply

1. Click 'Apply' on our portal and register or log in.
2. Complete the screening form and upload your resume.
3. Increase your chances of being shortlisted for an interview with the client.

About Uplers

Uplers is committed to making hiring processes reliable, simple, and fast. We support talents in finding and applying for relevant contractual and permanent opportunities, aiding their career progression. We are here to assist with any challenges faced during engagements.

Note: Additional opportunities are available on the portal. Successful assessments may qualify you for other roles.

If you are ready for a new challenge, a dynamic work environment, and an opportunity to advance your career, apply today!