faizan naviwala

I build production-ready AI systems that stay dependable in real environments—especially through logging, schema design, and safe tool orchestration. I like working in small, repeatable steps with clear, testable implementations.

As an AI Engineer (Oct 2025–Present) at Spiderweb Technologies, I built a robust data ingestion pipeline that downloads date-wise JSON records, normalizes them into CSV, and persists into MySQL for auditability. I implemented a controlled two-agent architecture where one agent retrieves and refines context and a second agent generates user-facing responses using tool-calling patterns.

I integrated database-backed retrieval and vector search to reduce hallucinations and improve factual accuracy. I also designed logging, retries, and error-reporting so the pipeline is reproducible and debuggable in production, and collaborated with cross-functional teams to deploy AI components and provide integration support.

Earlier, I interned as a Data Analyst Intern at ERP Boss Solutions, where I developed a barcode generator integrated into the ERP and built a customer support automation bot (Respond.io) to route queries and improve response efficiency. I also produced Power BI reports and ran targeted messaging campaigns; in a Milestone internship, I used Python for data cleaning, preprocessing, and end-to-end automation projects.