GenAI AI/ML Test Engineer: Evaluation & Benchmarking

Improving; Hinoba-an; On-site; PHP 796,000 - 1,194,000; Full time

Job summary

Improving is seeking an AI/ML Test Engineer – GenAI to design and automate evaluation strategies for Agentic AI applications in Mumbai. You will develop datasets, test cases, and benchmark suites to measure agent performance, reasoning quality, tool usage, and workflow effectiveness.

The role emphasizes identifying hallucinations, biases, safety risks, and failure patterns while collaborating with AI Engineers and product teams to improve reliability and specify actionable recommendations.

Qualifications

• Strong experience in Generative AI, LLMs, and Agentic AI systems.
• Hands-on expertise with AI evaluation frameworks (RAGAS, DeepEval, TruLens, LangSmith, Promptfoo, etc.)
• Proficiency in Python and AI/ML development libraries.
• Knowledge of Prompt Engineering, prompt testing, and optimization.
• Ability to define and track evaluation metrics such as accuracy, relevance, groundedness, hallucination rate, latency, and user satisfaction.
• Experience in creating automated evaluation pipelines and benchmarking frameworks.
• Strong understanding of AI safety, guardrails, bias testing, and responsible AI practices.
• Familiarity with REST APIs, JSON, vector databases, and knowledge retrieval systems.
• Experience in A/B testing, human-in-the-loop evaluation, and red teaming.
• Strong experience in Manual Testing of AI/GenAI applications, including functional, exploratory, UAT, regression, and end-to-end testing.
• Expertise in validating Agent Reasoning, Tool Calling, Workflow Execution, and Response Quality.
• Hands-on experience in Automation Testing using Python frameworks.

Responsibilities

• Design, execute, and automate evaluation strategies for Agentic AI applications.
• Develop evaluation datasets, test cases, and benchmark suites.
• Measure and improve agent performance, reasoning quality, tool usage, and workflow effectiveness.
• Analyze model outputs and identify hallucinations, biases, safety risks, and failure patterns.
• Collaborate with AI Engineers, Product Teams, and Domain Experts to improve agent quality and reliability.
• Generate evaluation reports, dashboards, and actionable recommendations.

Skills

Generative AI

LLMs

Agentic AI systems

AI evaluation frameworks

Python

Prompt engineering

Evaluation metrics

Automation testing

Agent reasoning

Tool calling

Workflow execution

REST APIs

JSON

Vector databases

A/B testing

Human-in-the-loop evaluation

Red teaming

Quality assessment

Tools

RAGAS

DeepEval

TruLens

LangSmith

Promptfoo

Job description

Improving is seeking an AI/ML Test Engineer – GenAI to design and automate evaluation strategies for Agentic AI applications in Mumbai. You will develop datasets, test cases, and benchmark suites to measure agent performance, reasoning quality, tool usage, and workflow effectiveness.

The role emphasizes identifying hallucinations, biases, safety risks, and failure patterns while collaborating with AI Engineers and product teams to improve reliability and specify actionable recommendations.