J
Staff Machine Learning Engineer
Actively Hiring
Full-time Posted about 1 month ago
Responsibilities
- check_circle Design, train, and optimize domain-specific ASR models for aviation and operational communications.
- check_circle Develop verticalized speech models tuned for specialized terminology, accents, abbreviations, call signs, and noisy radio/audio conditions.
- check_circle Build and maintain large-scale transcription and labeling pipelines for supervised and semi-supervised learning workflows.
- check_circle Fine-tune foundation speech models (e.g., Whisper, wav2vec, Conformer, RNN-T, Citrinet, NeMo-based architectures) for aviation-specific use cases.
- check_circle Improve transcription quality through language model adaptation, pronunciation lexicons, contextual biasing, and decoding optimization.
- check_circle Develop evaluation frameworks and benchmarking methodologies using WER, CER, domain entity accuracy, latency, and robustness metrics.
- check_circle Collaborate with product, avionics, data engineering, and platform teams to deploy scalable real-time and batch transcription systems.
- check_circle Optimize inference pipelines for edge, cloud, and low-latency streaming environments.
- check_circle Research emerging techniques in speech enhancement, diarization, speaker adaptation, multilingual ASR, and audio foundation models.
- check_circle Ensure compliance with security, privacy, and operational reliability standards required in aviation environments.
Basic qualifications
- Bachelor’s or Master’s degree in Computer Science, Machine Learning, Electrical Engineering, Linguistics, or related field.
- 3+ years of experience in speech recognition, audio ML, or applied machine learning.
- Strong experience training and fine-tuning ASR models using frameworks such as PyTorch or TensorFlow.
- Experience with modern ASR architectures including:
- + Transformer-based ASR Conformer RNN-T CTC-based systems Encoder-decoder speech models
- Conformer
- RNN-T
- CTC-based systems
- Encoder-decoder speech models
- Experience working with:
- + Speech/audio preprocessing Forced alignment Language model adaptation Beam search decoding Noise robustness techniques
- Forced alignment
- Language model adaptation
- Beam search decoding
- Noise robustness techniques
- Familiarity with NVIDIA NeMo, Kaldi, ESPnet, Hugging Face, Whisper, DeepSpeed, or equivalent ecosystems.
- Strong Python engineering skills and experience building production ML systems.
- Experience with cloud infrastructure and ML deployment workflows (AWS, Kubernetes, Docker, CI/CD).
- Ability to work with large audio datasets and distributed training environments.
Preferred qualifications
- Experience building ASR systems for aviation, air traffic control, public safety, defense, or other mission-critical domains.
- Familiarity with VHF/UHF radio communications and noisy-channel audio processing.
- Experience with multilingual or code-switching ASR systems.
- Background in speech enhancement, keyword spotting, diarization, or speaker verification.
- Knowledge of LLM-assisted transcription correction and retrieval-augmented speech systems.
- Experience optimizing models for real-time streaming inference.
- Active pilot experience or familiarity with aviation operations is a plus.
- Medical, dental, vision insurance with Employer paid health premiums
- Open PTO Policy
- 401(k) with up to 10% company matching and immediate vesting
- 12 Weeks Paid Maternity Leave
- 4 Weeks Paid Paternity Leave
- Flight Training Rewards
Tags & Focus Areas
Remote Machine Learning Data Engineer Ai