Senior AI/ML Engineer

At Quest Global, it’s not just what we do but how and why we do it that makes us different. With over 25 years as an engineering services provider, we believe in the power of doing things differently to make the impossible possible. Our people are driven by the desire to make the world a better place—to make a positive difference that contributes to a brighter future. We bring together technologies and industries, alongside the contributions of diverse individuals who are empowered by an intentional workplace culture, to solve problems better and faster.

About the Role

We are building AssureDroid AI , an autonomous AI-driven platform that validates Android applications across the OS ecosystem. You will be a core engineer on the Agentic Layer — designing and shipping the multi-agent system (Orchestrator, UI Understanding, Action Execution, Regression Detection, Root Cause Analysis) that powers the platform.

You will design production-grade ReAct/Planner-Executor agent loops, build the LLM router that abstracts Gemini/OpenAI/Anthropic, implement perceptual-hash caching that drives our cost economics, and own the prompt engineering discipline that determines whether a regression is correctly detected or missed.

Role Expectation: 

Design and implement core agents (UI Understanding, Action Execution, Regression Detection, Root Cause Analysis) using Gemini 1.5 Pro and Flash

Build the LLM Router with multi-provider abstraction, fallback chains, and cost aware routing

Develop the three-layer guardrails system (pre-LLM input sanitization, post-LLM schema validation, runtime safety)

Engineer the Cost Governor — real-time budget tracking, automatic Pro→Flash switching, hard-stop enforcement

Author and version production prompt templates with A/B testing harness

Build the Replay Engine for decision trace logging and run reproducibility

Drive the Eval & Learning Loop — golden test sets, prompt regression testing, fine-tuning signal collection

Must-Have Skills

6+ years software engineering, with 1+ years on production LLM/agentic systems

Strong Python (3.10+), with deep familiarity in FastAPI, asyncio, structured logging

Hands-on experience with Gemini, OpenAI, or Anthropic APIs — including streaming, tool use, structured outputs

Production experience with agent frameworks (LangChain, LangGraph, AutoGen, or custom orchestration)

Demonstrable understanding of prompt engineering as a discipline — versioning, A/B testing, golden sets

Experience with vector databases (FAISS, Pinecone, Vertex Matching Engine) and embedding pipelines

Solid grasp of distributed systems patterns — message queues, event-driven architecture, idempotency

Nice to Have

Experience with Temporal, Cloud Workflows, or stateful workflow engines

Background in Android internals, ADB, or mobile testing tooling

Published research, blog posts, or open-source contributions in agentic AI

Exposure to fine-tuning, RAG architectures, or model evaluation pipelines

GCP / Vertex AI hands-on experience

What You’ll Bring (Beyond Code)

Pragmatism over hype — you know when NOT to use an LLM

Cost-consciousness — you instinctively tag every LLM call with metadata

Strong written communication (decision docs, prompt versioning rationale)

We are known for our extraordinary people who make the impossible possible every day. Questians are driven by hunger, humility, and aspiration. We believe that our company culture is the key to our ability to make a true difference in every industry we reach. Our teams regularly invest time and dedicated effort into internal culture work, ensuring that all voices are heard.

We wholeheartedly believe in the diversity of thought that comes with fostering a culture rooted in respect, where everyone belongs, is valued, and feels inspired to share their ideas. We know embracing our unique differences makes us better, and that solving the worlds hardest engineering problems requires diverse ideas, perspectives, and backgrounds. We shine the brightest when we tap into the many dimensions that thrive across over 21,000 difference-makers in our workplace.