Principal Software Developer, Autodesk

• Bachelor's or Master's degree in Computer Science, Machine Learning, Data Science, or equivalent practical experience
• 8+ years building cloud-native software in production (distributed systems, APIs, data-intensive services, reliability and operations)
• 4+ years delivering AI/ML-powered system in production
• Traditional ML cloud applications (training pipelines, deployment, monitoring, iteration)
• LLM-based systems (RAG, MCP, Agent workflows, fine-tuned models)
• Experience with MCP or similar standardized patterns for connecting models to tools and context
• Experience with deploying and maintaining AI Applications in production reliably, monitoring performance, and improving over time
• Proficiency in Python/TypeScript/Java with strong engineering fundamentals (testing, code quality, performance, security)
• Strong communication skills: you can explain tradeoffs clearly and influence decisions without relying on authority

Desirable

• The Applied AI team in Autodesk's Data and Process Management (DPM) organization ships Cloud-Native AI agents that make our Product Data Management (PDM) and Product Lifecycle Management (PLM) workflows smarter and easier
• This is a hands-on architecture role where you'll design core platform patterns and also write code that lands in production. We're looking for a modern builder: Ambitious, Curious, and Practical—someone who has already shipped scalable, Cloud-Native AI Applications to production and wants to level up further
• You might have expertise in GenAI Applications, or you might have shipped traditional Machine Learning applications (recommendations, forecasting, anomaly detection) at scale and are ready to go all-in on agentic architectures. Either way, you build for reliability, quality, and impact
• Architect and build scalable, secure cloud-native services and Agentic AI workflows that run in production
• Own the GenAI/ML architecture for production agentic systems: tool-use, orchestration, state/memory, routing, and multi-step workflows
• Define model strategy across prompting, retrieval (RAG), and fine-tuning—making tradeoffs using measurable quality, latency, safety, and cost
• Standardize tool/context integrations across internal systems using MCP-based patterns (or equivalent approaches), enabling teams to ship faster on a shared foundation
• Establish evaluation + observability standards (regression tests, monitoring, feedback loops)
• Set technical direction via reference architectures, best practices, and hands-on technical guidance across teams
• Partner closely with Product, Security/Privacy, and Engineering leaders to deliver high-impact features