Hi, I'm Rahul Jeyasingh

I'm an engineer by profession, a thinker by nature.

Data engineer with 4+ years of experience in Python, SQL, and GCP. I spent roughly three years at Infosys on an onsite assignment in Berlin (September 2022 through July 2025), where I built production data systems — most notably a config-driven similarity matching microservice that became the most reusable piece of infrastructure I've designed. I also presented a CAN bus security tool at Black Hat Arsenal 2022.

Python, SQL, and GCP aren't resume keywords — they're the tools I actually think in.

Calculating time lived...

My MBTI personality type is INTP.

What Drives Me

As a child, I folded origami, rode bicycles, and broke my toys — not out of carelessness, but to understand what was inside. I played with toy motors and basic electronics. I derived math and geometry results independently, working through proofs not because a teacher assigned them but because I wanted to know why the formulas worked.

The games tell the same story: RollerCoaster Tycoon (systems design), Age of Empires 2 (resource management under constraints), Cities: Skylines (infrastructure engineering).

This curiosity-first wiring hasn't changed. The tools evolved — from toy motors to ESP32 boards, from breaking toys to building interpreters — but the impulse is the same.

What I'm Working On

  • ▸DRISHTI MDL #70 — AI-based design change impact analysis for Mazagon Dock Shipbuilders. Neo4j, fine-tuned Mistral 7B, and the similarity matching system from Infosys form the backbone.
  • ▸Five-year horizon — Financial freedom, building something of my own, location independence, and control over my time.

Intellectual Interests

My deepest preoccupation is how intelligence and life emerge from physical laws. I've independently designed theoretical systems for evolving intelligence based on photonic principles — photon tollgates as logic gates, information compression through feature connections, and unified architectures for real-time learning and inference.

This connects to Karl Friston's Free Energy Principle, photonic computing, and predictive coding — frameworks I arrived at overlapping conclusions with before encountering the literature.

On the AI side, I've studied LLM architecture at a foundational level — working through Andrej Karpathy's pure-Python GPT implementation, building intuition for tokenization, autograd, embeddings, and attention. I've explored multi-agent frameworks (LangGraph, CrewAI, Claude Flow, AutoGen) and can distinguish between open-weight and open-source models — a distinction many practitioners gloss over.

Hardware hasn't been left behind: Arduino and ESP32 experiments, Raspberry Pi robotics (maze-solving, line-following competitions), and building an assembly-like language from scratch — a constant itch to work at layers most software engineers never touch.

The throughline from breaking toys to building production systems to designing photonic intelligence frameworks is a single thread: learn by building, build by being curious, and don't wait for permission or a use case.