Hey, I'm Aditya,
I am a software engineer and systems researcher working at the intersection of applied AI, agent loops, and distributed backend infrastructure. At Synopsys, I design internal engineering tools , build low-latency dashboards processing realtime CI/CD metrics. I also build several automation agents and SKILLS reducing development time by 75%.
My open-source work centers on solving agent fragility. I created AdiClaw, an autonomous AI CLI coding assistant featuring a deterministic staging pipeline that isolates and verifies code modifications before disk mutation. I also built Studio, a native macOS utility performing local Whisper ML transcription directly on Apple Silicon.
Through Coding Adda, I have mentored over 15,000 developers in modern software engineering and generative AI, and authored the handbook Basics of Prompt Engineering: A Practical Guide.
Architecture & Engineering Tenets
How I approach software design, autonomous agent loops, and systems engineering.
01. Determinism Over Stochastics
Direct filesystem edits by AI agents cause cascading errors and silent regressions. Every code mutation must be isolated in a staging environment and verified against automated test suites before disk commit.
02. First-Principles Reconstruction
True engineering depth comes from stripping away magic. I understand databases, KV-cache indexing, and multi-agent coordination by rebuilding their core algorithms from scratch in plain code.
03. Evaluation-Driven Engineering
Prompts and agent system instructions are software artifacts. Never optimize behavior based on anecdotal runs. We establish reproducible benchmarks, synthetic evaluation harnesses, and regression gates.
Basics of Prompt Engineering: A Practical Guide
A hands-on, engineering-driven guide bridging the gap between intuitive experimentation and production-grade AI systems. Written to help developers structure, evaluate, and benchmark system instructions for real-world reliability.
