Raj Chhapariya
WorkAboutWriting
ResumeContactGitHub
Raj Chhapariya•© 2026•Bengaluru, India•Privacy
GitHubX (Twitter)LinkedInEmail
Technical Publications & Systems Research

Writing

Thorough technical explorations covering information retrieval, safe autonomous agents, vectorized database internals, modern web architectures, and LLM evaluation frameworks.

01
Aug 10, 2026
7 min read

Hybrid Retrieval Systems in Production: Combining BM25, Dense Embeddings, and Reciprocal Rank Fusion

A deep architectural dive into lexical search limitations, vector space blindspots, and the mechanics of Reciprocal Rank Fusion.

Pure vector search frequently fails on exact keyword identifiers and domain jargon, while lexical BM25 misses semantic paraphrasing. This article breaks down the mathematical mechanics and implementation of Reciprocal Rank Fusion (RRF) for production RAG pipelines.

RAGInformation RetrievalBM25Vector SearchPython
Read Deep Dive
02
Aug 14, 2026
5 min read

Deterministic Guardrails for LLM Agents: Enforcing Safe SQL and Pydantic Schemas without Open-Ended Code Execution

Why unrestricted Python code generation is an unacceptable security hazard, and how to build sandboxed analytical agents with AST parsing and numerical verification.

Giving language models unbounded code execution in analytical pipelines opens catastrophic security and hallucination vectors. This guide demonstrates how to architect deterministic, sandboxed agent tools using AST-level SQL inspection, DuckDB read-only boundaries, and Pydantic runtime schema contracts.

AI SecurityDuckDBPythonPydanticSQL
Read Deep Dive
03
Aug 18, 2026
5 min read

In-Process Columnar OLAP with DuckDB: Architecture, Vectorized Execution, and Analytics Engineering

An analysis of vectorized execution engines, Apache Arrow memory interoperability, and why in-process analytics outperforms client-server databases for local data workflows.

Traditional client-server databases introduce significant serialization overhead for analytical workloads. This deep-dive examines how DuckDB leverages columnar storage, Morsel-driven parallelism, and vectorized SIMD execution to process millions of rows directly in-process with sub-second response times.

DuckDBOLAPDatabase InternalsData EngineeringPerformance
Read Deep Dive
04
Aug 22, 2026
5 min read

Next.js 16 App Router Architecture: Server Components, Streaming SSR, and Static Site Generation

A technical exploration of React Server Component execution boundaries, serialization protocols, and runtime memory optimization.

The transition to React Server Components fundamentally altered how web applications manage state, bundle boundaries, and data fetching. This deep-dive examines Next.js 16 App Router mechanics, Partial Prerendering, RSC serialization contracts, and cache revalidation strategies.

Next.jsReactTypeScriptWeb ArchitecturePerformance
Read Deep Dive
05
Aug 25, 2026
4 min read

Evaluating Hallucination and Citation Faithfulness in Retrieval-Augmented Generation

Methodologies for measuring precision, citation recall, context contamination, and automated offline evaluation suites.

Building a RAG pipeline is straightforward; systematically evaluating its faithfulness and hallucination rate is where most engineering teams struggle. This article details automated evaluation frameworks, grounded assertion metrics, and cost-effective offline testing with mocked embeddings.

AI EvaluationRAGHallucinationTestingPython
Read Deep Dive