Rahul Rangarao

Summary

Analytics professional at Bristol Myers Squibb with an end-to-end data skill set, transitioning into AI engineering on the strength of shipped, production-quality open source. Author of knowledge-worker, a published PyPI package that gives AI assistants durable, provenance-backed memory: a local-first knowledge graph with graph analytics, LLM-backed ingest, and a hard provenance invariant. Generalist by nature: comfortable owning a problem from data ingestion through modeling, tooling, and the report on a stakeholder's desk.

Open Source

knowledge-worker · Author & Maintainer

• Designed and shipped a provenance-first knowledge graph engine: every durable claim must carry a source document and literal excerpt before entering the graph, enforced by deterministic validation.
• Built the full CLI surface: ingest, query, path-finding, review/merge, context export, deep-dive pre-ingest workspaces, and an offline HTML graph visualizer.
• Implemented a graph-analytics audit layer (PageRank, betweenness, k-core, community detection, weak-claim queues) so users can see what their AI memory actually knows, and where it is thin.
• Architected open-web storage: canonical JSON-LD, append-only JSONL history, Turtle/RDF interchange; zero runtime dependencies in the core, with LLM backends (Anthropic, OpenAI, Ollama) as optional extras.
• Established a 9-benchmark public test suite guarding provenance, recall, privacy boundaries, and context compactness; GitHub Actions CI across Python 3.10–3.13.
• Published to PyPI (v0.8.0); ~500 downloads/month and adopted for real cross-session AI workflows.

• kompress-ultra · context-management middleware experiments for AI agent frameworks: live-data TUI dashboards, knowledge-graph visualizations, Pages deploy pipeline.
• AG-UI protocol · open protocol work for agent/UI interoperability.

Experience

Bristol Myers Squibb · Analytics Professional

• Delivered end-to-end analysis, from data ingestion to stakeholder-ready reporting, within a single day; recognized with the Impact Award (Sept 2021).
• Recognized with a SPOT Award (Aug 2020) for stakeholder communication and building business context around analytics deliverables.

Education

The University of Texas at Austin · 2022 – 2023

Skills

Languages & Data

Python, SQL, Snowflake, data warehousing, JSON-LD / RDF / linked data

AI Engineering

LLM API integration (Anthropic, OpenAI, Ollama), agent context management, retrieval & memory systems, prompt/context engineering

Graph & Analysis

knowledge graphs, PageRank / betweenness / k-core / community detection, provenance modeling

Engineering Practice

CLI tooling, Python packaging (PyPI), GitHub Actions CI, benchmark-driven testing, local-first architecture

Certifications & Awards

• Hands On Essentials: Data Warehouse · Snowflake, Oct 2022
• Impact Award (2021) · SPOT Award (2020) · Bristol Myers Squibb
• Programming for Everybody (Python) · Coursera, 2016

Languages: English · Kannada · Hindi