AI & Automation Engineer specializing in Agentic Workflows, LLM Ops, and Autonomous Testing Frameworks.
Passionate about engineering reliable developer tooling, protocol-driven AI agent architectures (MCP), and resilient test automation with Playwright.
- 🤖 Agentic AI & LLM Ops: Designing autonomous multi-agent systems, protocol-driven developer tooling (MCP), and rigorous LLM evaluation pipelines using DeepEval and RAGAS to stress-test retrieval accuracy, hallucination rates, and tool-calling reliability.
- 🧪 Autonomous Testing & Diagnostics: Building next-generation web automation engines, diagnostic parsers, and self-healing test frameworks with Playwright and PyTest.
- 🎓 Academic Background: B.Tech in CSE (Artificial Intelligence) from Amrita Vishwa Vidyapeetham (Rank: 6th | CGPA: 9.11/10).
- 📜 Published Researcher: 2 peer-reviewed conference publications in IEEE proceedings on cybersecurity botnet detection (AICAI 2024) and on-device code translation (ICCCNT @ IIT Mandi).
| Project | Category | Description | Key Tech |
|---|---|---|---|
| playwright-trace-doctor | QA & AI Diagnostics | The Instant MRI for Broken Web Tests. One-shot diagnostic & self-healing intelligence engine for Playwright traces with correlated upstream 5xx/4xx failure detection (<5ms latency, dual CLI & MCP). | Playwright TypeScript MCP Node.js CI/CD |
| mcp-verify | LLM Ops & Safety | ESLint for AI Tools. Automated LLM-readiness, schema contract, and safety linter for Model Context Protocol (MCP) servers, preventing agent hallucinations and unguarded mutations before deployment. | TypeScript MCP JSON Schema Ajv CLI |
| antigravity-skill-porter | Agentic Workflows | Bridge the AI Agent ecosystem: Transpiler and optimizer porting Claude Code & Cursor skills into Google Antigravity native multi-agent plugins with parallel subagents (invoke_subagent). |
Python Agentic AI Google Antigravity AST |
| omniget | Systems & Tooling | Universal Windows package manager wrapper unifying WinGet, Chocolatey, and Scoop with parallel Runspaces, live environment variable broadcasting, and declarative sync. Packaged on WinGet (Nexus.OmniGet). |
PowerShell WinGet Windows API CLI |
| HTTP C2 Detection & Analysis | Published Research | Real-time behavioral detection of HTTP-based Command & Control botnets using MongoDB and machine learning. Published in IEEE AICAI 2024. | Python MongoDB Scikit-Learn Threat Intel |
| On-Device Code Translation | Published Research | Lightweight AST-based on-device semantic code translation between dynamically typed interpreted languages. Published in IEEE ICCCNT at IIT Mandi. | Python AST Compilers Language Translation |
| TRASIS | FinTech & Trading | Algorithmic trading bot combining moving average crossover strategies (SMA/WMA) and financial news sentiment analysis for automated trade execution. | Python Pandas Sentiment Analysis NumPy |
AI & LLM Ops : Model Context Protocol (MCP), Agentic Workflows, DeepEval, RAGAS,
LangChain, Prompt Engineering, Computer Vision, OCR (AWS Textract, Tesseract)
Automation & QA : Playwright, Selenium, PyTest, AI Monitoring Protocols, Regression Testing,
End-to-End Test Framework Architecture, Self-Healing Locators
Languages : Python, TypeScript, JavaScript, SQL, C++, Java, PowerShell, HTML/CSS
Tools, Cloud & CI/CD : AWS (EC2, S3, Lambda), GCP, Docker, GitHub Actions, MongoDB,
Mixpanel, Appcues, Git
I take pull requests and code review seriously:
- Clean Architecture & Reproducibility: Every tool includes clear setup commands, automated test suites, and strict linting.
- Contract & Schema Safety: Enforcing deterministic schema contracts for agentic tools and robust error handling.
- Automated Verification: Standardized GitHub Actions CI workflows on every repository to prevent regressions.
- Contributions Welcome: All flagship repositories include PR templates, issue templates, and comprehensive
CONTRIBUTING.mdguides.
Let's build reliable, agentic systems together. Feel free to connect via LinkedIn or reach out at pranav.03.h@gmail.com.