Artificial Intelligence Intern

Role Overview

Neurohaze AI is seeking a Voice AI Intern to help build and improve production-grade conversational voice agents. You will work across the complete voice pipeline, including speech recognition, dialogue logic, LLM reasoning, tool calling, and speech synthesis. This is a hands-on role focused on turning prototypes into dependable, real-world systems.

Responsibilities

• Build and iterate on voice-agent pipelines involving streaming ASR/STT, intent detection, dialogue logic, LLM reasoning, and TTS.
• Integrate speech and language APIs from different providers.
• Evaluate providers based on accuracy, latency, reliability, and cost.
• Develop prompts, tool definitions, and conversation flows.
• Profile the end-to-end pipeline to identify bottlenecks and reduce latency.
• Test voice systems against challenging audio conditions like accents and background noise.
• Create evaluation harnesses and automated test suites.
• Support Arabic-language voice capabilities, specifically Gulf dialects.
• Document systems, experiments, and deployment processes.

Requirements

• Currently pursuing or recently completed a degree in Computer Science, Artificial Intelligence, Machine Learning, Data Science, or related field.
• Strong Python programming skills, including asynchronous code and REST APIs.
• Working knowledge of LLM concepts such as prompting, context management, and tool calling.
• Familiarity with ASR/STT, TTS, or audio processing.
• Ability to read technical documentation and integrate unfamiliar APIs.
• Clear written and spoken English.
• Availability for a three-month remote internship starting between 2 August and 6 September 2026.

Nice to Have

• Experience with AI voiceover, Generative AI tools, or LLM evaluation.
• Knowledge of LLMOps, web development, or Arabic NLP.

Skills

• Python
• LLM
• ASR/STT
• TTS
• REST APIs

Benefits

• Internship completion certificate
• Letter of recommendation
• Five-day work week
• Hands-on experience building real-world voice and generative AI systems