Senior NLP & Platform Engineer

Company Description

NeuraNx.ai is a cutting-edge technology company focused on creating AI-powered products and innovative cloud solutions for businesses ready to lead in the AI-driven era. Specializing in areas such as LLMs, RAG systems, multi-agent platforms, and cloud architectures across Azure, AWS, and GCP, NeuraNx.ai enables automation, scalability, and operational confidence. With a team of seasoned practitioners who bring real-world expertise across industries like automotive, aviation, finance, and energy, NeuraNx.ai is committed to delivering practical, impactful, and efficient technology solutions. Guided by a mission to make intelligent technology accessible, the company defines itself as a partner in shaping the future of enterprises eager to embrace AI.

About the role

You own everything between raw ASR output and the final structured legal transcript plus the production infrastructure to serve it. This role combines language model development (domain LMs, dialogue classification, Q&A formatting) with MLOps (model serving, API, review UI, monitoring). You are the bridge between model quality and deployable software.

What you'll do

NLP & post-processing (60%):

• Build domain 4-gram KenLM from 500M+ words of legal text
• Fine-tune DeBERTa-v3-base for dialogue act classification (QUESTION/ANSWER/OBJECTION/RULING)
• Implement Q&A formatter with legal transcript conventions (EXAMINATION BY headers, direct/cross/redirect tracking)
• Build punctuation restoration, inverse text normalization, and legal entity normalization
• Implement hot-word biasing logic (collaborate with Speech Engineer)

Platform & MLOps (40%):

• Build FastAPI backend for model serving
• Deploy models using Docker, orchestrate with Redis/Celery for batch processing
• Build web-based review UI (or manage a frontend contractor) with audio playback and confidence highlighting
• Set up monitoring, logging, and alerting
• Ensure fully offline, air-gapped deployment (zero external calls)

What you bring

• 5+ years in NLP engineering and experience deploying ML systems
• Experience training language models (n-gram and/or neural)
• Hands-on fine-tuning of transformer classifiers (BERT/DeBERTa family)
• Experience with FastAPI, Docker, and model serving (vLLM, Triton, or similar)
• Ability to build a functional web UI or manage a frontend contractor
• Strong Python engineering and system design

Nice to have

• KenLM, OpenGrm, or other n-gram LM toolkits
• Legal text experience (citations, statutes, formatting)
• React/Vue for review UI
• Prior speech/audio ML deployment
• Redis, Celery, or job queue experience