itssujeeth Selected Projects
Outcome-first write-ups on what was built, why, and what shipped. Each entry states its status honestly: a prototype is labelled a prototype, and a demonstration that never reached production says so.
Featured
A deterministic-first procurement automation platform for a precious metals retailer, automating reorder decisions across four metals and five sales channels with AI constrained strictly to an advisory role.
An MLOps and LLMOps platform for a Fortune 500 automotive aftermarket retailer, originated with a working demo, won in April 2025, approved by the architecture review board, and carrying 9 of 27 models in production by September 2026.
A five-zone enterprise agentic-AI platform that lets non-technical business teams self-serve AI-generated insights from governed enterprise data across six business domains.
A GenAI-assisted approach to remediating known security vulnerabilities across an enterprise codebase, budgeted at a fraction of the manual estimate and in progress when I left.
The enterprise generative AI blueprint for a large logistics company, answered with working proofs of concept across Google and Microsoft stacks rather than slideware. None of the POCs went to production, and I say so.
A single command that points Claude Code at any inference backend (local Ollama, Anthropic, OpenAI, or Gemini) with a clean environment and no manual env-var juggling.
A 24-week, self-paced interactive curriculum shipped as a React app that teaches IT support managers to use Claude across their actual daily work.
A prototype conversational analytics agent that lets leaders ask natural-language questions of millions of short free-text field notes, with every number computed by tools and every answer linked to its evidence.
Archive
An end-to-end proof-of-concept running ML model monitoring across three deployment paradigms (custom sklearn, AutoML Tabular, and BigQuery ML) on Google Vertex AI with Terraform-provisioned infrastructure.
A production-ready template that collapses the first 3 hours of MCP server setup to a single clone.
Cut the regression detection cycle from days to a 4-minute CI gate for LLM-based agents.
An LLM-assisted approach that extracted, summarized, and correlated requirements from a large volume of recorded discovery sessions inside an enterprise modernization program, on a schedule manual analysis could not meet.
Classification and entity-extraction models that cut a logistics brokerage's email bid response time from about 32 minutes to about 4, with the analyst confirming every extraction before a price went out.
An image-processing automation that eliminated a recurring high-volume manual workload, saving ~1,000 staff hours per year.
A static-code-analysis system that maps a sprawling enterprise application portfolio using knowledge graphs and graph neural networks, surfacing consolidation candidates data-driven rather than by intuition.
A computer-vision demonstration that reads product attributes from packaging imagery, built to show how manual catalog tagging could be automated.
A custom text-recognition pipeline for reading item details from product packaging where off-the-shelf OCR failed on custom fonts. Delivered as a demo and roadmap; not deployed.
A source-code scanner that found every application affected by an enterprise-wide store-numbering change across a dozen languages, and became the seed of the portfolio rationalization product.
Hiring systems for one of the world's largest retailers: a QR-coded document-attachment design the client chose over a planned manual team, and a later background-check cost redesign projected to save about $4M a year.