ML & LLM Ops Software Engineer

Role Description We are seeking an experienced ML & LLM Ops Engineer to own the operational backbone of our ML and GenAI platform, taking models from experimentation to reliable, production-grade systems and keeping them healthy over time. This is a hands-on role for someone who thrives in a startup and can make ML and GenAI run reliably at scale, including in the on-prem and edge environments common to industrial customers. You will: Own the end-to-end operational lifecycle of ML and LLM systems. Build and maintain automated CI/CD pipelines for model training, deployment, and serving. Run quality monitoring, drift detection, and observability for models in production. Operate LLM Ops workflows: prompt versioning, evaluation, guardrails, and inference optimization. Ensure reliability, security, and reproducibility using best-in-class ML Ops and LLM Ops practices. Must Have: Bachelor's degree in engineering or higher, plus 3+ years in ML Ops, ML platform, or ML infrastructure engineering. Strong experience deploying, serving, and maintaining ML models in production. Hands-on experience with MLflow and CI/CD automation for ML. Proficiency with Docker and Kubernetes/K3s. Experience with major cloud infrastructure across Azure, GCP, and AWS, and with Databricks. Experience operating LLM applications (inference serving, evaluation, guardrails) and model quality monitoring. Strong software engineering fundamentals, ideally in Python. Ownership mindset: self-motivated, adaptable, and driven to take projects from concept to production. Preferred to have: Experience deploying open source LLMs (e.g., Llama, Mistral, Qwen) to on-prem or edge machines. Experience with inference serving frameworks such as vLLM, TGI, Triton, or Ollama. Experience running models in air-gapped or resource-constrained environments. Experience operating ML systems for real-time time series data (sensor/IoT). About Spector.ai Spector.ai is a well funded early-stage startup solving the $1.5 trillion challenge of industrial asset reliability. We are building an AI-first industrial agent platform that moves plant reliability from reactive to autonomous operations, combining machine learning with domain-specific AI Agents to deliver real-time diagnostics, root cause analysis, and actionable recommendations at scale. With early pilots and major industry partnerships in motion, we are pioneering the future of AI-powered plant health.