Automatic Speech Recognition Engineer

About the role

The Automatic Speech Recognition (ASR) Engineer will join Omilia's Delivery ASR team to own the end‑to‑end development, tuning, testing, deployment and live performance of ASR technology for enterprise IVR and voice‑channel solutions. This position requires strong technical judgment, the ability to work under tight deadlines, and direct interaction with customers and partners.

Key responsibilities

• Own end‑to‑end ASR tuning, model and grammar delivery for live enterprise IVR/voice deployments, from requirements through go‑live and ongoing optimization.
• Investigate root‑cause of production ASR issues and ensure rapid recovery of customer‑facing voice channels.
• Maintain the technical relationship for assigned enterprise accounts as the primary ASR delivery contact.
• Drive measurable WER/SER improvements through acoustic and language model design, including modern neural approaches.
• Build and scale delivery automation pipelines to reduce manual turnaround.
• Monitor live ASR performance across global regions and proactively address degradation.
• Deploy and validate ASR updates across production infrastructure.
• Document tuning decisions, model versions and configuration changes for auditability.

Required profile

• 5+ years of hands‑on ASR/speech engineering experience, preferably with production IVR or voice‑channel deployments.
• Proven track record supporting customers directly, ideally in regulated sectors such as banking, insurance or healthcare.
• Strong programming ability (Python preferred; Bash, Java or C++ acceptable) and comfort working in Linux environments.
• Academic background in a relevant technical field.

Required skills

• Experience with ASR platforms such as Kaldi (language modeling, acoustic modeling, grammar/lexicon design, decoder tuning).
• Acoustic and language model design, including modern neural network approaches.
• Python programming.
• Bash, Java or C++.
• Linux command‑line environment.