Ensylon interview questions

The discussion started with a brief introduction about my background and recent projects. The interviewer then quickly moved into a **deep dive on my Deloitte project**, asking detailed questions around: * RAG pipeline design * Handling large PDF documents (10–80 MB, multi-page parsing) * LLM chaining and orchestration * Failure handling (retries, fallback, validation layers) * State management in multi-step workflows They were particularly interested in **production challenges**, like: * Handling LLM hallucinations * Ensuring deterministic outputs * Latency and scaling (1000 concurrent requests scenario) * Async vs queue-based architectures --- * **Design thinking** * Clean structure * Handling edge cases --- ### 👍 **What I Liked** * Interview was **very practical and scenario-based** * Focused on **real production challenges**, not textbook questions * Interviewer was knowledgeable and pushed for **depth of understanding** * Good discussion around **AI system architecture** * Writing clean Python code for workflows * Explaining system design clearly * **Difficulty Level:** Medium to High * **Focus Area:** Applied GenAI + Backend Engineering * **Recommendation:** Great opportunity for engineers with hands-on experience in **LLMs, RAG, and production AI systems** Overall, a **great learning experience** and a solid interview process for anyone serious about working on real-world AI systems.