AI/ML Lead (Speech, Model Tuning & Guardrails)

PeopleTec is currently seeking an AI/ML Lead (Speech, Model Tuning & Guardrails) to support our Huntsville, AL location.

PeopleTec is seeking an AI/ML Lead in Huntsville, AL, to develop speech recognition, model-tuning, and safety capabilities for an on-premise AI system supporting the H-1 flight simulator. You will work with AI/ML, software, and aviation subject matter experts to adapt models to cockpit audio and military terminology, evaluate their performance, and ensure model outputs are safely handled before they can affect the simulator. 

• Develop and fine-tune speech and language models for cockpit audio, military terminology, and simulator use. 
• Build and maintain training and evaluation data, including testing with speakers and data not used for training. 
• Develop and test safeguards that validate AI outputs before they can affect simulator actions, and document the results. 
• Prepare models, data, and evaluation materials for delivery and work with the team on continued model improvement. 

Required Skills/Experience:

• Preferred experience range: 5–8 years of machine learning engineering experience, including development and delivery of language-model applications to users. Demonstrated technical and leadership ability will be considered regardless of total years of experience. 
• Hands-on experience fine-tuning speech recognition models for domain-specific terminology and evaluating performance using held-out data. 
• Strong Python skills and experience serving models on-premise using CUDA-capable hardware. 
• Experience building and evaluating a retrieval-augmented generation (RAG) system that produced reliable, expected results. 
• Ability to travel occasionally for test events. 
• Must be a U.S. Citizen

• Bachelor’s degree in computer science, mathematics, engineering, or a related field preferred. Equivalent relevant professional or military experience may substitute for a degree. 

• Experience integrating speech recognition and text-to-speech, including voice activity detection and audio pipelines for noisy environments. 
• Experience with structured or schema-constrained generation, model-serving adapters, speech/audio pipelines, knowledge graphs or ontologies, and AI applications for aviation, training, or simulation. 
• Experience with knowledge graphs and ontologies (OWL, RDF, SPARQL) and rule-based validation of generated structures. 
• Experience in aviation, tactical training, or simulation, and working alongside an operator subject matter expert.