AI Engineer - Voice AI & Agents (GCP ADK)

Mira Labs Ai; Pune District; On-site; INR 4,000,000 - 7,000,000; Full time

Benefits offered by this job

Relocation allowed

On-site Pune office

Job summary

Mira Labs Ai in Pune is hiring AI Engineers to build the AI core of our platform: production voice AI agents and multi-agent workflows for enterprise lenders. The role emphasizes hands-on work with Google Cloud ADK and Vertex Gemini for real-time voice agents and latency-aware design.

You will design, implement, and optimize agent systems, mentor peers, and collaborate across cloud providers to deliver robust mortgage conversation solutions.

Qualifications

• 5–8 years of software/ML engineering experience, with 2+ years building LLM-powered products in production.
• Hands-on experience with Google Cloud ADK (Agent Development Kit) and the Vertex AI / Gemini ecosystem.
• Proven experience building and shipping Voice AI agents — streaming speech-to-text/text-to-speech and real-time pipelines.
• Telephony/real-time audio experience: Twilio, LiveKit, SIP, WebRTC, or comparable.
• Strong Python engineering fundamentals — production-grade code, not just notebooks.
• Deep practical understanding of agent design: tool calling, planning, guardrails, and failure handling.
• Experience designing evals and observability for LLM/agent systems.

Responsibilities

• Design and build production agentic systems using GCP ADK and Vertex AI (Gemini).
• Develop real-time Voice AI agents: streaming STT/TTS, turn-taking, and telephony integration.
• Architect multi-agent orchestration: tool calling, function routing, memory, guardrails, and handoff.
• Build RAG pipelines over mortgage documents with grounding and citation behavior.
• Own quality: build eval harnesses, regression suites, and monitoring for non-deterministic agent behavior.
• Optimize for production: latency, cost, token budgets, caching, and fallback strategies across providers.
• Work with Azure OpenAI alongside GCP — multi-cloud model usage.
• Mentor engineers and set patterns for how the team builds with LLMs.

Skills

Python

LLM systems

GCP ADK

Vertex AI

Voice processing

Telephony

Tools

ADK (Agent Development Kit)

Twilio

LiveKit

SIP

WebRTC

Vertex AI

Job description

The Role

We are hiring AI Engineers to build the AI core of our platform: production voice AI agents and multi-agent workflows that handle real mortgage conversations and processes for enterprise lenders across North America. This role demands hands‑on depth with Google Cloud's Agent Development Kit (ADK) and real-time voice agent development — not just API calls, but latency budgets, barge‑in handling, tool orchestration, and evals.

Location: Onsite (Pune)

Working Hours: 7.30 pm to 4.30 am

What You'll Do

• Design and build production agentic systems using GCP Agent Development Kit (ADK) and Vertex AI (Gemini)
• Develop real‑time Voice AI agents: streaming STT/TTS, turn‑taking, interruption/barge‑in handling, and telephony integration (Twilio/SIP)
• Architect multi‑agent orchestration: tool calling, function routing, memory, guardrails, and human handoff
• Build RAG pipelines over mortgage documents and guidelines with strong grounding and citation behavior
• Own quality: build eval harnesses, regression suites, and monitoring for non‑deterministic agent behavior
• Optimize for production: latency, cost, token budgets, caching, and fallback strategies across model providers
• Work with Azure OpenAI alongside GCP — we are pragmatic and multi‑cloud on models
• Mentor engineers and set patterns for how the team builds with LLMs **

Must-Have Qualifications

• 5–8 years of software/ML engineering experience, with 2+ years building LLM‑powered products in production
• Hands‑on experience with Google Cloud ADK (Agent Development Kit) and the Vertex AI / Gemini ecosystem
• Proven experience building and shipping Voice AI agents — streaming speech‑to‑text/text‑to‑speech (Google Cloud Speech, Deepgram, ElevenLabs, or similar) and real‑time conversational pipelines
• Telephony/real‑time audio experience: Twilio, LiveKit, SIP, WebRTC, or comparable
• Strong Python engineering fundamentals — you write production code, not just notebooks
• Deep practical understanding of agent design: tool calling, planning, state, guardrails, and failure handling
• Experience designing evals and observability for LLM/agent systems

Nice to Have

• Experience with Azure OpenAI, LangGraph, or other agent frameworks
• Fine‑tuning or model‑optimization experience
• Exposure to regulated domains (fintech, lending, healthcare) and compliance constraints on AI systems
• Contributions to open‑source AI tooling

Compensation

No equity

Job Location

Pune

Remote Work Policy

In office

Visa Sponsorship

Not Available

Preferred Timezones

Central Time

Relocation

Allowed

Skills

• Azure
• GCP
• ADK

Hiring contact

Suketu Gaglani

CEO 3 years