Hey — I'm Adi.

Applied AI engineer.

Adi Vardi
Machine Learning · Quant · Systems · Deep Learning · FullStackMachine Learning · Quant · Systems · Deep Learning · FullStack

About

I

am

an

applied

AI

engineer

working

at

the

intersection

of

machine

learning

computational

modeling

and

production

systems.

I

build

end

to

end

AI

infrastructure

research

modeling

algorithmic

design

backend

architecture

automation

pipelines

and

deployment

in

demanding

multidisciplinary

R&D

environments.

I

care

about

scientific

rigor

operational

reliability

and

long

horizon

systems

thinking

especially

for

physics

driven

data

intensive

problems.

Autodidactic collaborative and growth driven excellence distinctions from military and Ministry of Education. Open to collaborations where ambitious engineering meets real world impact.

Courses

DeepLearning.AI
Amazon Web Services

Data Engineering Professional Certificate

AirflowSparkAWSTerraformSQL
DeepLearning.AI

Deep Learning Specialization

TensorFlowCNNsRNNsHuggingFacePython
Stanford Sign
Stanford
DeepLearning.AI

Machine Learning Specialization

PythonNumPyScikit-learnTensorFlowNeural Nets
DeepLearning.AI

PyTorch for Deep Learning Professional Certificate

PyTorchCNNsTransfer LearningHuggingFaceTransformers

Skills

JavaScript
TypeScript
React
Python
C#
Git and GitHub
HTML
CSS
SQL
Machine learning
Deep learning
Amazon Web Services
Google Cloud Platform
Linux

Intra stack

TensorFlow
PyTorch
OpenCV
Scikit-learn
Supabase
Vercel
Railway
PostgreSQL
Apache Kafka
TradingView
Claude (Anthropic)
Cursor
CrewAI
Hugging Face

Ask me anything

Ask about Adi
Ask about Adi’s work, stack, or how to connect — answers stay grounded in what’s on this site.

The trip

From data foundations to defense-scale AI

Four legs — analyst work, hardened backends, quant product, and mission-critical applied ML.

  • Now

    Applied AI Engineer

    Israeli Military Intelligence · INI

    Tel Aviv · on-site

    Elite multidisciplinary R&D: AI-driven computational systems for physics-heavy domains, end-to-end data pipelines, ML/DL in performance-critical simulation-adjacent stacks. INI Commander’s Excellence Distinction (2026).

  • Parallel

    Founder · Trader

    Quantrade

    Remote

    Quant equity & macro work plus Quantrade: private portfolio optimization SaaS — ML recommenders, risk-parity allocation, Monte Carlo engines, systematic construction grounded in modern portfolio theory.

  • 2022

    Backend Python Developer

    SheeldS (Arilou Automotive Cybersecurity)

    Ramat Gan · part-time

    Backend R&D in automotive cybersecurity: engineered a recursive Python hex parser decoding low-level CAN telemetry into structured data models. Built Python & C# utilities for log normalization, signal extraction, and troubleshooting workflow automation. Operated within ASPICE & ISO 21434-compliant cross-functional processes.

  • 2021

    Data Analyst

    Nebius Israel

    Tel Aviv · part-time

    High-volume web data acquisition at Nebius (ex-Yandex): extracted and structured large-scale external datasets into reliable analysis-ready assets. Delivered actionable insights through advanced analytics. Automated recurring collection pipelines for operational scale.

Selected work

Quant platform, defense-scale ML/CFD stacks, and production backends — links where public.

Quantrade

Quantitative finance platform — portfolio optimization SaaS with ML-driven signals, risk-aware allocation, and simulation-backed workflows. Live product.

Quantrade preview
SkyGuard-ML

SkyGuardML is a real-time detection and tracking system optimized for tiny aerial objects, combining a custom YOLO11 architecture with advanced tracking and post-processing to achieve high-accuracy performance on challenging micro-targets. It integrates detection, tracking, analytics, and visualization into a unified pipeline for scalable, production-grade aerial surveillance.Built for edge deployment and mission-critical counter-UAS workflows, with sub-meter micro-target tracking under real-world latency constraints.

SkyGuard-ML preview
WaveR

WaveR is a high-resolution ocean forecasting and routing engine that combines real-time wind, wave & advanced metrics data with interactive visualization to support marine decision-making. It enables users to analyze forecasts at any point or plan optimized multi-leg routes with dynamic conditions and ETA calculations.Turns live wind, wave, and ocean telemetry into actionable routing intelligence for operators navigating shifting marine conditions.

WaveR preview
TerraFlow

TerraFlow is a data-to-simulation pipeline that converts real-world terrain and building data into CFD-ready environments for accurate water and wind flow analysis. It automates the full workflow—from geospatial ingestion to simulation outputs—enabling scalable, physics-based environmental modeling.Bridges raw geospatial inputs and CFD-ready meshes so teams can run water and wind flow analysis without manual prep overhead.

TerraFlow preview
Automotive cyber telemetry

ISO 21434-aligned CAN parsing, structured threat models, and backend automation for fleet-scale detection platforms (SheeldS / Arilou).

Automotive cyber telemetry preview

More detail in conversation.

Let’s build something sharp

Best first touch is LinkedIn — or email if you prefer a longer brief.

adivardi2005@gmail.com