Physics + Engineering + AI

Building AI for the
Physical World.

IIT Kanpur · Physics & Aerospace Engineering.
I build production AI systems that reason about physical reality — solar energy infrastructure, sensor telemetry, simulation engines, and agentic intelligence over real-world data.

See the Work Get in Touch
profile.json
// physics-trained engineer
{
  "name": "Mubashshir",
  "foundation": [
    "Physics", "Aerospace Eng"
  ],
  "alma_mater": "IIT Kanpur",
  "builds": "AI × Physical Systems",
  "current": "ScadaNXT · Solar SCADA",
  "domains": [
    "Physics-Informed ML",
    "LLM Agents",
    "Energy Infrastructure",
    "Time-Series Forecasting"
  ],
  "interest_in": [
    "General Relativity",
    "AGI", "Quantum Mechanics"
  ],
  "open_to": true
}
G_μν + Λg_μν = (8πG/c⁴) T_μν
About

Where Physics Meets
Production Systems

I hold a double major in Physics and Aerospace Engineering from IIT Kanpur — one of India's top technical institutions. But my trajectory wasn't toward academia. It was toward the question: what happens when you put first-principles physical thinking inside production software?

"Most engineers simulate. I model the underlying physics and let the simulation follow."

As Founding Software Engineer at TerraNxt, I built ScadaNXT — a solar SCADA and operational intelligence platform — from the ground up. That meant writing the inverter physics models, the power flow engine, the LLM agents querying live telemetry, and the infrastructure that stitches it all together.

I'm interested in roles where the physics isn't decorative — where understanding degradation curves, thermal dynamics, plasma behavior, or orbital mechanics is the actual engineering edge.

Education
B.Tech · Physics + Aerospace Engineering
IIT Kanpur
Current Role
Founding Software Engineer
TerraNxt · ScadaNXT Platform
Previous
Research Scientist · Lead Product Engineer
Smarttrak AI
Location
Gurgaon, India · Open to Relocation
US · EU · Remote
Deep Interests
General Relativity · Quantum Mechanics · AGI
Mubashshir - Professional portrait
Technical Stack

What I Build With

The intersection of physical modeling, production systems, and machine intelligence.

Physics & Simulation

First-principles modeling is the foundation. I write the physical equations before I write the code.

pvlib Solar Geometry Power Flow Electrical Graph Synthetic Data Gen GR · QM

AI & Agentic Systems

LLM agents that reason over real sensor data. Not chatbots — reasoning engines grounded in physical telemetry.

LangGraph Multi-Agent RAG Time-Series ML Physics-Informed ML InfluxDB

Infrastructure & Systems

Production-grade backend that handles real-time data, async pipelines, and multi-user concurrency.

FastAPI PostgreSQL SQLAlchemy WebSockets Docker / AKS MinIO

Frontend & Visualization

Interfaces built for operators who need physical insight, not dashboards that look good in demos.

React Three.js React Flow Digital Twin 3D Plant Builder Grafana

Computer Vision & Remote Sensing

Processing satellite and drone imagery for physical asset inspection and mapping.

SAM 2 WorldView Rooftop Detection RTX 4060 Async Pipelines

Mathematical Reasoning

Comfortable with the deep math underlying physical systems — from information theory to differential geometry.

Tensor Calculus Info Theory Kolmogorov Complexity Combinatorics Group Theory
Experience

Where I've Shipped

2023 — Present
Founding Software Engineer
TerraNxt · ScadaNXT Platform · Gurgaon, India
Built the entire ScadaNXT solar SCADA and operational intelligence platform from scratch. Authored physics-based synthetic data generation, a power flow engine with electrical graph traversal, a five-node LangGraph multi-agent system over InfluxDB telemetry, a 3D digital twin plant builder in Three.js, and a React Flow topology editor. Sole engineer responsible from architecture to deployment.
FastAPILangGraphpvlibInfluxDBReactThree.jsPostgreSQL
Full Stack
2021 — 2023
Research Scientist · Lead Product Engineer
Smarttrak AI
Research and product engineering at an AI company. Combined scientific research methodology with production engineering — a pattern that defines the current trajectory.
Research + Eng
Projects

Things Built From First Principles

Production systems where physical understanding was the engineering advantage.

Flagship · Production

ScadaNXT — Solar SCADA & Operational Intelligence

A full B2B solar monitoring and intelligence platform for utility-scale and rooftop solar assets. The platform ingests real sensor telemetry, models degradation against physical baselines, and exposes a conversational LLM interface for plant operators to query their own data in natural language.

Architecture: LangGraph multi-agent graph (router → dashboard → fetcher → analyst → synthesizer) over live InfluxDB time-series data, via a FastAPI WebSocket backend and a React frontend.

LangGraphFastAPIPhysics ModelingInfluxDBReactWebSocketDocker
Physics Engine
PR = E_gen / (G_poa × P_nom / G_stc)
Custom solar geometry model
θ_z = arccos(cos φ cos δ cos ω + sin φ sin δ)
Electrical graph traversal
P_loss = Σ I²R across {strings, combiners, inverters}
Inverter efficiency surface
η(P, T) = f(P_dc/P_dcₒ, T_amb, V_dc)
Computer Vision · Geospatial

Rooftop Detection Pipeline

Satellite imagery segmentation using SAM 2 on WorldView commercial imagery to detect and measure rooftop solar potential at scale. GPU-accelerated inference on RTX 4060 Laptop with async processing pipeline feeding PostgreSQL.

SAM 2WorldViewRTX 4060GeospatialAsync
SaaS · Production

Mighty Resume

Production SaaS (mightyresume.in) with GPT-4o-mini ATS scoring for resumes. Deployed on Azure Kubernetes Service with GitHub Actions CI/CD, cert-manager TLS, and nginx ingress. Full engineering ownership from code to cloud.

AKSCI/CDGPT-4o-mininginxcert-manager
3D Visualization

Digital Twin Plant Builder

3D solar plant builder in Three.js with drag-and-drop asset placement, real-time PostgreSQL mutations via SQLAlchemy, and a React Flow topology editor for electrical graph configuration. Built for operators to model their physical plants digitally before commissioning.

Three.jsReact FlowPostgreSQLDigital Twin
Intellectual Interests

How I Think About the World

Physics isn't a credential — it's an epistemic framework. These are the domains that shape how I approach engineering problems.

General Relativity & Curved Spacetime

The Einstein field equations as a constraint on geometry. I'm comfortable with the stress-energy tensor, gravitoelectromagnetic formulations, and the Aichelburg–Sexl ultraboost. Physics at this depth changes how you think about approximations in engineering.

AGI & the Physics of Intelligence

Scaling laws as empirical physics of a new kind. The question of whether intelligence emergence is substrate-neutral — and what that implies for how we build systems — sits at the intersection of information theory and long-horizon civilizational thinking.

Information Theory & Complexity

Kolmogorov complexity, combinatorial entropy via group theory, and the deep connection between compression and physical law. Information-theoretic reasoning is underused in engineering and overused in hype — I try to tell the difference.

Open to Opportunities

Let's Build Something
Real.

I'm actively looking for roles at the intersection of physical systems, AI, and genuine engineering ambition — frontier labs, deep-tech companies, energy/aerospace infrastructure. If you're working on something that requires understanding the physics, not just the API, reach out.

Email GitHub LinkedIn Resume Twitter / X