◈ HOUSE OPERATIVE // ARRAKIS SECTOR // YEAR 10191

ADITYA SINGH

AI ENGINEER · FULL-STACK · ML SYSTEMS

I build AI-powered products that innovate. Driven by love of the game — I work with precision, adapt relentlessly, and don't stop until the vision is exactly what it needs to be.

[ VIEW THE MISSIONS ]
[ GITHUB ARCHIVE ↗ ]
SYS.BOOT SEQ
LAT: 28.4°N · LONG: 14.7°E · SPICE LEVEL: CRITICAL
STATUS: AVAILABLE FOR MISSIONS_
TYPESCRIPT◈
PYTHON◈
REACT / NEXT.JS◈
NODE.JS / EXPRESS◈
FASTAPI◈
MACHINE LEARNING◈
DEEP LEARNING / CNNs◈
DIFFUSION MODELS◈
LLM INTEGRATION◈
AI AGENTS◈
SYSTEM DESIGN◈
POSTGRESQL / MONGODB◈
BLOCKCHAIN / WEB3◈
DOCKER◈
TYPESCRIPT◈
PYTHON◈
REACT / NEXT.JS◈
NODE.JS / EXPRESS◈
FASTAPI◈
MACHINE LEARNING◈
DEEP LEARNING / CNNs◈
DIFFUSION MODELS◈
LLM INTEGRATION◈
AI AGENTS◈
SYSTEM DESIGN◈
POSTGRESQL / MONGODB◈
BLOCKCHAIN / WEB3◈
DOCKER◈
◈ SIETCH RECORD // MODULE: OPERATIVE_PROFILE

I Work Until
The Vision Is Real.

"The mystery of life isn't a problem to solve, but a reality to experience."

— DUNE · FRANK HERBERT

I'm a full-stack engineer and AI builder — focused on systems that actually matter. Not demos. Not experiments. Products with real architecture and real stakes behind them.

What drives me is love of the game. I work with precision, I adapt fast, and I don't ship until I've overpowered the problem. Every project is a chance to mold a vision exactly in my image.

I'm targeting the kind of work that pushes limits — AI integrations, ML systems, full-stack platforms. If the problem is hard and the stakes are real, I'm interested.

OPEN FOR: AI INTEGRATIONS · FULL-STACK · API DEV · ML ENGINEERING
// CORE READOUT — 4 VITAL METRICS
PRIMARY
AI SYSTEMS
TERTIARY
FULL-STACK WEB
COMPUTE STRATEGY
EDGE & CLOUD
DATA ARCH
RELATIONAL & VECTOR
SYSTEM LOG: The architecture prioritises resilience. Heavy adoption of serverless combined with custom inference instances. Model fine-tuning is conducted via low-rank adaptation on constrained hardware to guarantee deployment efficiency.
◈ CHOAM REGISTRY // MODULE: MISSION_LOG

SELECTED OPERATIONS.

OP-01 // CLASSIFIED[ AI · ML ]

RETINAI

Multi-disease retinal classification using diffusion model-based anomaly maps.
A medical AI system that detects multiple retinal diseases from fundus images. Built on a diffusion model pipeline for anomaly map generation, paired with specialised CNNs per disease class and a meta-classifier that aggregates outputs for final prediction. Designed for real diagnostic utility — not a toy model.
PYTHON·PYTORCH·DIFFUSION MODELS·CNN·META-CLASSIFIER
>> ACCESS ARCHIVE ↗
OP-02 // ACTIVE[ AI AGENT ]

BUDAI

Real-time intelligent AI agent — makes calls, manages tasks, operates autonomously.
A live AI agent that operates in real-time — capable of making calls, managing tasks, and acting as an always-on personal assistant. Takes action autonomously.
AI AGENT·VOICE·LLM
>> ACCESS ARCHIVE ↗
OP-03 // DEPLOYED[ WEB3 ]

GLADIATOR ARENA

On-chain PvP game — win fights, claim opponents' weapons.
A fully on-chain PvP battle game with real economic stakes. P2P architecture with smart contract enforcement.
BLOCKCHAIN·SMART CONTRACTS
>> ACCESS ARCHIVE ↗
OP-04 // OPERATIONAL[ ML SYSTEMS ]

FRAUDEYE

Hybrid ML + rule-based fraud detection API.
A system combining machine learning models with deterministic rule engines — capturing both pattern-based anomalies and hard logic.
PYTHON·FASTAPI·SCIKIT-LEARN
>> ACCESS ARCHIVE ↗
OP-05 // DEPLOYED[ FULL-STACK ]

HANGOUT-HERO

AI itinerary generator — real-time plans built on mood, location, and group preferences.
Generates personalised real-time hangout itineraries by combining user mood, live location data, group size, and preferences. Powered by OpenAI with live maps integration.
TYPESCRIPT·OPENAI·MAPS API·NEXT.JS
>> ACCESS ARCHIVE ↗
// NODE CONFDISTRIBUTED
// LATENCY14MS BASE
// BANDWIDTHNOMINAL
// STATUSACTIVE
◈ SPICE ARCHIVES // MODULE: FIELD_CHRONICLES

DISPATCHES FROM
THE DEEP DESERT.

AI RESEARCHSUBSTACKMEDIUM
12 MIN READ

The Ego Gate — Why AI Needs to Learn What Not to Learn

Intelligence without a filter is not intelligence. Every AI system ever trained lives closer to a vessel for noise than you think. The Ego Gate is a framework built on three principles — Doubt, Curiosity, and Arrogance — that teaches neural networks to selectively ignore, buffer, and dream. A new approach to the stability-plasticity dilemma.

"How do we make AI choose what not to learn? Because intelligence without a filter is a very expensive mirror."
READ FULL CHRONICLE CHR-001
FULL-STACKMEDIUM
5 MIN READ

Hangout Hero — Your AI Itinerary Generator

It started from a personal problem — we wanted an app that optimises weekend plans around mood, budget, and group preferences. So we built one. Powered by Gemini 2.5 Pro, it generates a personalised hangout itinerary in under 60 seconds.

READ FULL CHRONICLE CHR-002

MORE DISPATCHES
INCOMING

// NEW CHRONICLES IN TRANSMISSION
// SUBSCRIBE TO INTERCEPT

SUBSCRIBE ON SUBSTACK
READ ON MEDIUM
ARCHIVE STATUS: 2 CHRONICLES INDEXED · PLATFORMS: SUBSTACK + MEDIUM · CLEARANCE: PUBLIC_
◈ DEEP DESERT // MODULE: OPEN_CHANNEL

Transmit a Signal.

Available for AI integrations, full-stack web apps, API development, and ML engineering. If the problem is hard and the build is real — let's talk.

GITHUB ARCHIVE
EMAIL CHANNEL
LINKEDIN RELAY
"He who controls the spice controls the universe."
— DUNE · FRANK HERBERT