Hi! I applied for the Agentic AI & Automation Trainee position. Given the team’s focus on frameworks like LangGraph, Model Context Protocol (MCP), and tools like Claude Code or Cursor, what primary challenges are you currently solving when integrating AI
Yaswanth M. asked a question to Casey C.
- 1 reply
- 0 views
- Category: Role, Role challenges
- Date asked:
- Last update:
-
Hello Yaswanth M.,
Thanks for the question - glad you've already applied!
To be honest, I'm not directly on the Agentic AI and Automation team, so I can't speak for their exact day-to-day work. But from my side as a software architect, I'd share two challenges I see around integrating AI.
First, AI output can look good but still be wrong - someone still needs to verify it against the real requirement, spec, or system behavior. Sometimes it sounds so authoritative, but you still need an expert to check if it actually makes sense.
Second, technical design and architecture. Some features or implementations need to work with other ongoing features and implementations, and that's not always easy for AI to spot. Solutions and optimization often come from experts, technical leaders, or architects who understand the bigger system.
Tools can help a lot with the work - but you still need human expertise in the loop.
Good luck! Feel free to send me a follow-up question - and ping me once you're in!
Best regards,
Casey
