Coherent Solutions is hiring an AI/ML Developer (Speech AI)

Our client is a technology startup building advanced voice automation solutions for the quick-service restaurant industry. The company develops a privacy-conscious, high-performance drive-thru voice assistant that automates real-time customer interactions, improves order accuracy, and helps restaurant chains increase revenue and operational efficiency. Its product is designed for fast deployment in noisy, high-volume drive-thru environments and is already gaining market traction through cooperation with a major restaurant chain. The project is a next-generation voice automation engine for drive-thru order-taking in quick-service restaurants. It enables fully automated, real-time conversations between customers and the restaurant ordering system using modern speech recognition, natural language processing, and text-to-speech technologies. The engineer will work on improving the core voice and audio capabilities of the platform, including Speech-to-Text, Text-to-Speech, noise cancellation, speech enhancement, and real-time audio pipelines. The main focus will be on reducing latency, improving recognition quality and speech clarity, and making the system robust in extremely noisy real-world environments such as drive-thrus. Optimize low-latency, real-time Speech-to-Text pipelines for production drive-thru environments; Improve Text-to-Speech naturalness, responsiveness, and overall conversational quality; Design, tune, and improve noise suppression, echo cancellation, and speech enhancement systems; Improve speech recognition accuracy and robustness under challenging acoustic conditions, including engine noise, weather, overlapping speech, poor microphone quality, and outdoor environments; Build and scale audio processing infrastructure for production deployments; Evaluate, benchmark, and compare speech models using real-world audio data and production scenarios; Experiment with modern Speech AI technologies, models, and architectures to improve system performance; Collaborate with LLM and conversational AI teams to improve end-to-end voice interaction quality;