AI Automation Test Lead

Newbridge; Singapore; On-site; SGD 180,000 - 240,000; Full time

Job summary

Newbridge seeks an AI Automation Test Lead to shape quality strategy for AI/LLM products. You will build and scale test automation frameworks that validate model behavior, data quality, and non-deterministic outputs, leading a team of SDETs to ensure reliability, safety, and performance in production.

You will own end-to-end QA for chatbots, RAG pipelines, and model endpoints, implement eval harnesses, monitor metrics, and drive safety, reliability, and performance in production environments.

Qualifications

• 8+ years in software QA/test automation, with 2+ years leading teams.
• 3+ years hands‑on testing AI/ML systems, LLMs, or data‑intensive platforms.
• Strong coding in Python for test framework development. Java/Go is a plus
• Experience with test automation: Pytest, Playwright, Selenium, REST/GraphQL, CI/CD with GitHub Actions, Jenkins
• Deep understanding of LLM/RAG concepts: prompts, embeddings, vector DBs, chunking, eval metrics
• Hands‑on with Flink/Spark, SQL, Hive for validating data pipelines
• Experience with cloud + K8s: AWS/GCP, Docker, Kubernetes, model serving on GPU/CPU
• Built eval pipelines using LangSmith, Langfuse, Weights & Biases, MLflow, or similar
• Strong grasp of statistics for A/B testing, significance, and measuring non‑deterministic systems

Responsibilities

• Define and own the end-to-end QA strategy for AI/LLM products including chatbots, agents, RAG pipelines, recommendation, and CV/NLP models
• Establish best practices for testing non-deterministic systems: prompt evaluation, hallucination detection, bias/safety testing, latency & cost regression
• Lead, mentor, and grow a team of SDETs/automation engineers. Set quality gates for CI/CD and release readiness
• Partner with Product, Data Science, and Engineering to shift-left quality and define acceptance criteria for AI features
• Architect automation frameworks for AI systems: prompt regression suites, golden dataset evaluation, synthetic data generation, LLM-as-judge pipelines
• Build tooling to test RAG quality: context relevance, grounding, citation accuracy, retrieval latency
• Automate testing of model APIs, vector DBs, embedding pipelines, and fine-tuning workflows
• Implement eval harnesses using frameworks like DeepEval, RAGAS, LangSmith, Promptfoo, or custom solutions
• Design tests for data pipelines feeding AI: schema validation, drift detection, feature consistency between training/serving
• Own offline/online eval pipelines. Track metrics: accuracy, faithfulness, toxicity, P50/P95 latency, token cost
• Build canary & shadow testing for model deployments. Define rollback criteria based on guardrail violations
• Drive API, UI, and integration test automation for services hosting AI models
• Performance, load, and chaos testing for LLM inference endpoints and real-time features
• Security testing for prompt injection, jailbreak, data leakage, and PII handling
• Create quality dashboards: model eval trends, defect leakage, flaky rate, coverage for AI scenarios
• Drive root cause analysis for AI incidents. Feed learnings back into dataset curation and test design
• Ensure compliance with AI safety, privacy, and regulatory requirements

Skills

Python coding

Test automation leadership

A/B testing statistics

CI/CD

LLM concepts

Tools

Pytest

Playwright

Selenium

REST/GraphQL

GitHub Actions

Jenkins

LangSmith

Langfuse

Weighs & Biases

MLflow

Flink

Spark

SQL

Hive

Docker

Kubernetes

Job description

We are seeking an AI Automation Test Lead to drive the quality strategy for our clients AI/LLM-powered products. You will build and scale test automation frameworks that validate not just functionality, but also model behavior, data quality, and non-deterministic outputs. This role combines deep test automation expertise with hands‑on experience in AI/ML systems, leading a team to ensure our AI features are reliable, safe, and performant in production.

Key Responsibilities

1. Test Strategy & Leadership

• Define and own the end-to-end QA strategy for AI/LLM products including chatbots, agents, RAG pipelines, recommendation, and CV/NLP models
• Establish best practices for testing non-deterministic systems: prompt evaluation, hallucination detection, bias/safety testing, latency & cost regression
• Lead, mentor, and grow a team of SDETs/automation engineers. Set quality gates for CI/CD and release readiness
• Partner with Product, Data Science, and Engineering to shift-left quality and define acceptance criteria for AI features

2. AI Test Automation Architecture

• Architect automation frameworks for AI systems: prompt regression suites, golden dataset evaluation, synthetic data generation, LLM-as-judge pipelines
• Build tooling to test RAG quality: context relevance, grounding, citation accuracy, retrieval latency
• Automate testing of model APIs, vector DBs, embedding pipelines, and fine-tuning workflows
• Implement eval harnesses using frameworks like DeepEval, RAGAS, LangSmith, Promptfoo, or custom solutions

3. Data & Model Quality

• Design tests for data pipelines feeding AI: schema validation, drift detection, feature consistency between training/serving
• Own offline/online eval pipelines. Track metrics: accuracy, faithfulness, toxicity, P50/P95 latency, token cost
• Build canary & shadow testing for model deployments. Define rollback criteria based on guardrail violations

4. Traditional + AI System Testing

• Drive API, UI, and integration test automation for services hosting AI models
• Performance, load, and chaos testing for LLM inference endpoints and real-time features
• Security testing for prompt injection, jailbreak, data leakage, and PII handling

5. Governance & Reporting

• Create quality dashboards: model eval trends, defect leakage, flaky rate, coverage for AI scenarios
• Drive root cause analysis for AI incidents. Feed learnings back into dataset curation and test design
• Ensure compliance with AI safety, privacy, and regulatory requirements

Required Qualifications

• 8+ years in software QA/test automation, with 2+ years leading teams
• 3+ years hands‑on testing AI/ML systems, LLMs, or data‑intensive platforms
• Strong coding in Python for test framework development. Java/Go is a plus
• Experience with test automation: Pytest, Playwright, Selenium, REST/GraphQL, CI/CD with GitHub Actions, Jenkins
• Deep understanding of LLM/RAG concepts: prompts, embeddings, vector DBs, chunking, eval metrics
• Hands‑on with Flink/Spark, SQL, Hive for validating data pipelines
• Experience with cloud + K8s: AWS/GCP, Docker, Kubernetes, model serving on GPU/CPU
• Built eval pipelines using LangSmith, Langfuse, Weights & Biases, MLflow, or similar
• Strong grasp of statistics for A/B testing, significance, and measuring non‑deterministic systems

Preferred Qualifications

• Prior experience testing multi‑agent systems, tool use, function calling
• Knowledge of red‑teaming, AI safety evals, bias/fairness testing
• Contributions to open‑source AI eval or testing frameworks
• Experience with Doris, ClickHouse, Elasticsearch, Druid for test data analysis
• Background in FinTech, E-commerce, or Search domains with real‑time requirements