- LLM
- SQL
- AGENT
LLM SQL Agent
An LLM agent that answers questions over a SQL database: it translates natural language into queries, runs them read-only and summarises the result. Code at github.com/delcenjo/llm-sql-agent.
Plain-language questions over a SQL database, with no hand-written queries behind it: the agent is handed a few tools and left to inspect the schema, write a SELECT, do a bit of arithmetic and come back with an answer.
- Moderead-only
- Toolsschema + SQL
- DemoHF Spaces
Agency with limits
The agent can only run read-only queries, with bounded tools and a log of every step it takes. That is the lesson of the project: agency is useful exactly to the extent that it is well fenced.
Seeing the reasoning
Every answer arrives with the trace: which tables it looked at, what SQL it wrote and what the database returned. When the agent gets it wrong — and it does —, the trace turns the failure into something debuggable instead of a black box.
Try it live
The HuggingFace Spaces demo ships with a sample database: you ask the question, see the SQL it generates and the result. Watching the trace live explains the project better than any description.
LLM SQL Agent
Interactive demo hosted on HuggingFace Spaces
Next projects:
Credit Risk Platform
End-to-end MLOps platform for credit-default scoring: reproducible training with a versioned model registry, a FastAPI inference API with input validation, PSI drift monitoring and a Prometheus and Grafana observability stack, all containerised with CI. Code at github.com/delcenjo/credit-risk-platform.
Transformer from scratch
A GPT-style language model implemented from scratch in PyTorch: multi-head attention, causal masking and residual blocks written by hand, plus a byte-pair tokenizer and an ablation study. Code at github.com/delcenjo/transformer-from-scratch.
Built with
