661.3K
SHI started my career as a Software Engineer at Microsoft, working on a PaaS product on Azure. Solid engineering work, but I always knew I wanted to move closer to AI.
When I transitioned into working on an AI-agent based product in Copilot Studio, it wasn’t overnight or accidental.
I went back to the basics! Revised Python and ML fundamentals… then steadily built depth in GenAI: prompt engineering, embeddings, RAG, vector stores, and hands-on integrations with different LLMs.
I explored agent frameworks like LangChain and Semantic Kernel, and experimented with SmolAgents, CrewAI, and Bot Framework to understand different agent architectures and patterns.
Alongside this, I built small real-world projects like chatbots, retrieval systems, and automations… to see how these systems actually behave beyond demos.
I also focused on production-ready thinking: logging, monitoring, prompt versioning, and Responsible AI, because agents don’t live in notebooks… they run in real systems.
What made the transition effective wasn’t starting from scratch, but mapping my SWE experience to AI impact… using agents to automate workflows, improve search, reduce manual effort, and scale decision-making.
This isn’t a story about cracking interviews or having a PhD in AI. It’s about being consistent, practical, and curious while already being a software engineer.
I’ve put together a 6-month AI Engineer roadmap that you can follow (for both students and working professionals)
Comment “AI” and I’ll share the complete roadmap with you 📩
[ AI Engineer, AI Engineer Roadmap, Software developer, Software Engineer, Engineering, Microsoft, Story of my life, Career advice, Mentoring, Career guidance ]
#AI #Roadmap #Microsoft #SoftwareEngineer #Corporate
@shivanjaliverse
