Anam Ashraf

I’m a software product engineer focused on building scalable, AI-powered applications—especially LLM pipelines, semantic search, and RAG workflows. In my role as an SDE Intern at Xavier AI (Portugal, remote), I’ve worked in a fast-paced environment delivering high-concurrency systems with production-minded reliability.

I built a high-concurrency LLM data pipeline processing 17M+ records via SERP API translation and Groq batching, with Redis-based fault-tolerant state tracking and 24/7 VM orchestration. I also created 30+ AWS Lambda APIs to transform and route structured data between services, and optimized MySQL RDS to handle ~200 req/s through indexing, query refactoring, and batch inserts.

To improve retrieval and understanding, I embedded 60M+ records into Qdrant and optimized it using quantization for faster, more accurate vector retrieval. I’ve also Dockerized LibreOffice-headless for Word-to-PDF conversion, built FlexCel-based Excel-to-PDF with AI-assisted header identification (Claude-powered prompts) to preserve layout for ColPaLi ingestion, and developed tools like a custom lightweight fact-checker and an MCP-enabled research agent with a sidebar chat UI.