ZAZA
ABOUT / 01DISASSEMBLY REQUIRED

THIS IS THE“ABOUT ME” PART.

I’M SUPPOSED TO
EXPLAIN MYSELF.

COULD DO THAT.

OR WE COULD
TAKE IT APART.

ZAYAN ABOOBAKER.ZAZA, IF WE’RE BEING EFFICIENT. / B.E. ARTIFICIAL INTELLIGENCE & MACHINE LEARNING / 2025

OBJECT / ZA-01STATUS / OPENSCROLL TO DISASSEMBLE
02 / OPERATING SYSTEM
I DON’T HAND OFFTHE IDEA
TO SOMEONE ELSETO BUILD IT.
03 / TWO SIDESSAME BRAIN
ENGINEERING / 01

SYSTEMS
THAT THINK.

IIoT / SENSOR DATAANOMALY DETECTIONNLP / COMPUTER VISIONPYTHON / APIs
CREATIVE / 02

THINGS
YOU FEEL.

CINEMATOGRAPHYPHOTOGRAPHYGRAPHIC DESIGN1.2M+ CREATIVE VIEWS

DATA HAS NARRATIVE.

COMPOSITION HAS LOGIC.

THEY WERE NEVER
SEPARATE.
PRECISIONSTORY
04 / WHERE THAT CAME FROM01 / 06

I DIDN’T START

IN A BROWSER.

2021 — 2025THE FOUNDATION

I STARTED WITH
SYSTEMS THAT THINK.

B.E. in Artificial Intelligence & Machine Learning. Models, data, probability, Python — the part where correctness has consequences.

2024 — 2026UNIMATION ROBOTICS

THEN THE DATA
STARTED MOVING.

Industrial sensor streams, RS485 time-series pipelines, anomaly detection, predictive maintenance and real-time monitoring.

2024 — 2025ROOMAN / IBM / NASSCOM

TEXT. IMAGES.
MESSY INPUTS.

ML/NLP models, automated preprocessing, OCR and a medical diagnostic assistant. Less clean data. More real-world edges.

2025CDAC

THE MODEL NEEDED
AN INTERFACE.

An ATS-style system processing 500+ resumes. NLP extraction, ranking, Flask APIs — suddenly the way people used the system mattered too.

ALWAYSTHE OTHER SIDE

THE CAMERA WAS
ALREADY THERE.

Cinematography, photography, composition and timing. I was learning how to control attention long before I called it interaction design.

NOWCREATIVE DEVELOPMENT

EVENTUALLY,
THE BROWSER WON.

Not instead of engineering. Because of it. The system, the image, the motion and the implementation became the same problem.

05 / THE DETAILS PROBLEMI NOTICE THINGS

I NOTICE
THINGS.

THAT.

THAT TOO.

IT CAN GET
A LITTLE EXPENSIVE.

IN TIME, I MEAN.
06 / THE TECHNICAL BITS01 / 05

THE BORING

OFFICIAL BITS.

FINE.

HERE’S THE STACK.

Pick the problem. The stack reroutes itself.

01 / COMPUTER VISION

MAKE A MODEL
SEE.

Images in. Useful signal out. Classification, feature extraction, preprocessing and real-time inference.

02 / NLP + OCR

MAKE IT
READ.

Messy text, documents and prescriptions turned into structured input the rest of the system can actually use.

03 / ML SYSTEMS

MAKE IT
PREDICT.

Time-series, anomaly detection, feature engineering and model evaluation for data that changes in the real world.

04 / BACKEND + DEPLOYMENT

MAKE IT
RUN.

Models are only useful when something can call them. APIs, databases, testing and deployment glue the system together.

05 / AUTOMATION

MAKE THE
BORING PART
DISAPPEAR.

Scraping, browser automation and repeatable pipelines for the work nobody should be doing by hand twice.

LIVE ROUTINGSCROLL A PROBLEMHover any tool to isolate where it gets used.
LANGUAGES
PythonSQLJavaScriptCC++
ML / VISION
TensorFlowKerasPyTorchScikit-learnOpenCV
SPECIALTIES
Machine LearningDeep LearningNLPComputer VisionOCR PipelinesTime-SeriesAnomaly DetectionFeature EngineeringData PreprocessingModel OptimizationML Deployment
BACKEND / DATA
FastAPIFlaskREST APIsMySQLFirebase
TOOLS / AUTOMATION
GitGitHubJupyterPostmanSeleniumBeautifulSoupWeb Scraping
07 / THE HUMAN PART

I’M ZAYAN.

I learned to build models, APIs and systems before I started obsessing over what a browser could feel like.

The camera side was always there too — cinematography, photography, composition, timing. Eventually the two sides stopped pretending they were separate.

Now I make digital work where the idea and the implementation are the same job.

I’m still learning.
I intend to keep it that way.

08 / THE SEAMNEXT / PROJECTS
ENGINEERING
IMAGE / MOTION

ENOUGH ABOUT ME.

LOOK AT THE WORK.

THERE.