Voice AI Evaluation & Benchmarking Intern

Razer Inc.; Singapore; On-site; SGD 13,000 - 20,000; Part time

Job summary

Razer Inc. is seeking an Applied AI Intern to support AVA Voice AI, focusing on evaluating and improving voice interaction quality, latency and reliability.

You will work with AI and Software Engineers to benchmark models, investigate failure modes, and translate experimentation into production-ready improvements. Responsibilities include offline evaluation of voice AI models, implementing experiments to measure quality and latency, conducting error analyses, developing reusable evaluation

Qualifications

• Understanding of core AI/ML concepts including evaluation metrics, experimentation, and model benchmarking.
• Exposure to NLP, speech/voice AI, generative AI, or related fields.
• Strong programming skills in Python and SQL.
• Experience with backend application development using NodeJS or FastAPI is a plus.
• Knowledge of cloud technologies and containerization tools such as Docker or Kubernetes is a plus.
• Strong analytical and structured problem-solving skills, with curiosity and willingness to question assumptions.
• Excellent written and verbal communication skills.
• Ability to work effectively in a fast-paced, agile environment and learn through feedback.

Responsibilities

• Support offline evaluation and benchmarking of voice AI models, prompts, and pipelines using curated datasets.
• Implement experimentation workflows to measure quality, latency, reliability, and model performance.
• Conduct error analysis and edge-case investigations to identify failure modes and optimization opportunities.
• Develop reusable evaluation tools, benchmarking utilities, and automation to improve evaluation efficiency.
• Document evaluation findings and translate results into actionable recommendations for model and pipeline improvements.
• Conduct research on emerging AI, GenAI, and voice AI technologies to identify potential opportunities for adoption.

Skills

AI/ML concepts

NLP

Speech/Voice AI

Python

SQL

NodeJS

FastAPI

Docker

Kubernetes

Analytical thinking

Communication

Tools

NodeJS

FastAPI

Docker

Kubernetes

Job description

Razer Inc. is seeking an Applied AI Intern to support AVA Voice AI, focusing on evaluating and improving voice interaction quality, latency and reliability.

You will work with AI and Software Engineers to benchmark models, investigate failure modes, and translate experimentation into production-ready improvements. Responsibilities include offline evaluation of voice AI models, implementing experiments to measure quality and latency, conducting error analyses, developing reusable evaluation