Confirmed on the employer's own hiring board on Nov 17, 2026. First seen by Alion on Nov 16, 2026.
Develop and optimize speech-to-text systems that support Awarri’s multilingual applications, focusing on Nigerian languages. Design acoustic and language models, integrate them into real-time pipelines, and ensure robustness across noisy environments and varied accents.
Key Responsibilities
* Design, train, and fine tune ASR models such as Conformer, Transducer, CTC, or seq2seq.
* Build ASR pipelines that work for low-resource languages with limited labeled data.
* Optimize models for low latency, noise robustness, accented speech, and domain performance.
* Build scalable training and inference pipelines using PyTorch and GPU acceleration.
* Create data preprocessing pipelines for segmentation, normalization, augmentation, and labeling.
* Develop and maintain phoneme lexicons, G2P models, and MFA alignment workflows.
* Build evaluation suites for WER, CER, latency, rare-word recall, and accented speech.
* Diagnose failure modes such as mapping issues, poor alignment, or orthographic variation while handling orthographic inconsistencies and dialectal variations in underrepresented languages.
* Deploy ASR models to production with batching, streaming, and quantization optimizations.
* Monitor model drift, quality degradation, and real-time inference performance.
* Collaborate with product, linguistics, and ML teams to deliver production-ready ASR systems.
* Document experiments, model changes, and reproducibility workflows.
Person Profile
* Strong background in speech processing, signal processing, and deep learning.
* Hands-on experience training ASR models end-to-end.
* Skilled with PyTorch, GPU workflows, model optimization, and distributed training.
* Familiar with Conformer, RNN-T, CTC, seq2seq, and modern multilingual models.
* Comfortable with noisy, accented, or low-resource data environments.
* Experience with MFA, phoneme mapping, tokenization, and G2P tools.
* Strong debugging skills for alignment, segmentation, and data quality issues.
* Familiar with Whisper, MMS, wav2vec 2.0, HuBERT, and transfer learning.
* Able to write clean, production-ready code and build reliable ASR pipelines.
* Communicates clearly, flags blockers early, and owns tasks from idea to deployment.
Bonus: experience in African languages or other low resource languages

