Publications

From RAG to Agentic Workflows · Part 1 of 12

LLM Fundamentals in C#

A supplied judgment can be summarized only from passages you pass in. The same client, with no passage, names a citation that the catalog rejects.

From RAG to Agentic Workflows · Part 2 of 12

Simple RAG over Indian Court Judgments

Chunk a versioned judgment manifest, embed it locally, and return the passage with its character offsets.

From RAG to Agentic Workflows · Part 3 of 12

Advanced Legal Retrieval

Keyword, hashing-vector, and hybrid search on the same question, plus the SQL a pgvector store would run.

From RAG to Agentic Workflows · Part 4 of 12

Function Calling and Legal Research Tools

Deterministic C# tools for case metadata, passages, and a short constitutional catalog, with rejected blank input.

From RAG to Agentic Workflows · Part 5 of 12

An MCP Server in C#

A JSON-RPC handler exposes five read-only research tools and refuses any tool name outside that list.

From RAG to Agentic Workflows · Part 6 of 12

A Bounded Legal Research Agent

A tool loop with a call limit and a timeout that refuses outcome predictions and cites only retrieved passages.

From RAG to Agentic Workflows · Part 7 of 12

Agentic RAG and Precedent Research

Decompose a privacy question, keep a human-review flag, and do not treat a citation string as an overruling.

From RAG to Agentic Workflows · Part 8 of 12

A Multi-Agent Legal Research Workflow

Discovery, analysis, verification, and the report share one typed state and can resume after a stop.

From RAG to Agentic Workflows · Part 9 of 12

Event-Driven Judgment Ingestion

A local queue retries extraction, dead-letters repeated failures, and skips a job id that already completed.

From RAG to Agentic Workflows · Part 10 of 12

Agent Memory and Durable Execution

Save the research checkpoint separately from any chat transcript, then delete it.

From RAG to Agentic Workflows · Part 11 of 12

Evaluation, Security, and Reliability

Fixed checks for retrieval, unknown citations, hidden instructions in a judgment, and tools that are not allowed.

From RAG to Agentic Workflows · Part 12 of 12

A Local Legal Research Application

The ASP.NET Core API and Blazor UI that search the sample, run the workflow, and expose a health check.

Embeddings and Vector Databases: The Search Layer Behind Real AI Applications

Embeddings convert meaning into vectors; vector databases make those vectors searchable. Together they form the retrieval foundation behind RAG, semantic search, and enterprise AI assistants.

RAG vs Graph RAG vs Agentic RAG: How to Choose the Right Architecture

Standard RAG, Graph RAG, and Agentic RAG all connect LLMs to private data — but they solve different problems. Match the architecture to the question: retrieval, relationships, or reasoning.

Building Your First MCP Server in C# and .NET

Tutorial: expose product catalog search to AI hosts with the ModelContextProtocol NuGet package — compare manual HTTP integration vs MCP tools using enterprise domain models and Cursor setup.

MCP Tools, Resources, and Prompts in C#: A Hands-On Tutorial

Hands-on C# patterns for enterprise MCP servers — support ticket tools, customer profile resource URIs, documentation prompts, and contract testing with platform-domain examples.

Model Context Protocol Tools, Resources, and Prompts Explained

Understand MCP tools, resources, and prompts with practical examples for AI agents, enterprise workflows, and developer productivity.

Deploying MCP Servers: Stdio, HTTP, and Approval Gates

Platform engineering guide to MCP transport choice — stdio for Cursor, HTTP for production containers — plus human-in-the-loop approval before state-changing tools.