Agents · RAG · Machine Learning · Knowledge Graphs · Data Engineering
I'm an AI and Python engineer building practical intelligent systems: research agents, retrieval pipelines, graph ML, structured LLM workflows, and the data infrastructure behind them. I care about outputs that are explainable, testable, and useful beyond a demo.
Portfolio preview: all names, records, metrics, and scenarios shown above are synthetic. No real user or customer data is included.
| Project | What it demonstrates |
|---|---|
| GenAI Toolkit | A tested Python package and CLI for chat, streaming, Responses API calls, and tool loops across OpenAI-compatible providers. |
| Drug Interaction Discovery | A GNN, biomedical knowledge graph, and literature-retrieval pipeline for explainable drug-interaction research. |
| Real-Time Voting Pipeline | Kafka, Spark Streaming, PostgreSQL, Docker, and Streamlit working together in an end-to-end data engineering system. |
| Signal Desk Agent | An agentic research system that turns live web signals into typed intelligence briefs, with a credential-free demo and automated tests. |
| Ledger Tax Prototype | A live, dependency-free prototype for traceable AI tax workflows and transparent work prioritization. |
| CasePilot | An explainable legal document intelligence prototype with chronologies, issue spotting, evidence coverage, and citation-grounded work product. |
models PyTorch · GNNs · embeddings · structured generation
agents OpenAI APIs · tool calling · RAG · evaluation
python data pipelines · CLIs · automation · Streamlit
data PostgreSQL · Kafka · Spark · knowledge graphs
delivery Docker · AWS · APIs · CI/CD
- Multi-step AI agents with structured, inspectable outputs
- Retrieval and knowledge graphs for evidence-grounded reasoning
- Python systems that connect models, data, and real product workflows
Have an AI or Python problem worth solving?
akhilme008@gmail.com