Skip to content

Writing

Technical articles and thoughts

Breaking the Brittleness: How LLMs and VLMs Are Transforming UI Test Automation

September 25, 20257 min readPublished on ZySec AI Blog

Why traditional UI test automation is fragile, and how Large Language Models and Vision-Language Models enable natural language test generation and selector-free visual interaction.

aitestingautomationllm

Beyond Keywords: Architecting High-Fidelity RAG with Proactive Context

July 18, 20254 min readPublished on ZySec AI Blog

Shifting from reactive retrieval to proactive context engineering with Multi-Query Rewriting, HyDE, and GraphRAG for more reliable AI systems.

airagllmai-engineering

Processing 1B+ Records: Lessons from Building a Telecom Analytics Engine

Coming Soon~12 min

How we designed a distributed data pipeline with ClickHouse, Dask, and Prefect to process over a billion telecom records with sub-second query latency.

distributed-systemsdata-engineeringclickhouse

Building an AI Memory System: Smriti’s Architecture

Coming Soon~10 min

Why AI agents need persistent memory, how hybrid search works with vector embeddings, and lessons from writing 330+ tests for an MCP plugin.

aiopen-sourcetypescript

Multi-Agent Orchestration in Production: What I Learned Building C3

Coming Soon~8 min

Designing a SOAR platform with multi-agent orchestration, container isolation for security tools, and integrating 25+ providers.

aisecurityarchitecture