University Admissions Portfolio • Computer Science & Applied AI

Bridging mathematical vision,
educational equity, and distributed systems.

I am Ariet, an independent software engineer and student based in Bishkek, Kyrgyzstan. I build algorithms that solve real structural challenges: from high-precision computer vision for school exam evaluations to modern cloud infrastructures.

50k+
Exam Sheets Graded by Algorithm
99.8%
Subpixel OMR Accuracy
4
Fluent Languages (EN • RU • KY • TR)
500k+
Async Automated Tasks Orchestrated
“True engineering is not about complex syntax for its own sake. It is about taking mathematical principles — like homography, adaptive thresholds, and asynchronous event loops — and applying them to remove friction, bias, and inefficiency from everyday human lives.”
A
Ariet
Computer Science Candidate • Bishkek, Kyrgyzstan
English (Academic / IELTS Prep) Kyrgyz (Native) Russian (Fluent) Turkish (Fluent / Sapat Network)
Academic Background

Why Computer Science?

Growing up in Kyrgyzstan, I saw firsthand how educational institutions often lack expensive specialized equipment. While high-income countries deploy optical mark readers costing thousands of dollars, our regional schools relied on days of manual, error-prone teacher grading.

This drove my mission: to study Computer Science at an advanced collegiate level so I can develop software-defined solutions that replace prohibitive hardware with elegant algorithms accessible to schools and communities worldwide.

Cross-Disciplinary Dimension

Audio Engineering & Algorithmic Discipline

Beyond software, I have a deep creative passion for music production and acoustic signal engineering using Cubase and FL Studio.

Synthesizing sounds and balancing audio frequencies taught me the exact same virtues required in Computer Vision and distributed systems: harmonic balance, temporal precision, noise reduction, and rigorous attention to micro-details.

Flagship EdTech Engineering

AIM.EXAM: Democratizing National Exam Grading Through Computer Vision

A software-defined Optical Mark Recognition (OMR) system created to evaluate regional school olympiads and admissions tests without specialized hardware.

Applied Computer Vision De-facto standard in production

The Engineering Challenge & Solution

When teachers in regional schools scan stacks of thousands of exam sheets, the papers inevitably rotate by 5–15 degrees, shift by 10–20 pixels, or get crumpled. Standard OMR software simply misses the bubbles, causing false zeroes.

The Barrier

Conventional open-source OMR relies on rigid static bounding boxes that fail completely when sheets skew or scan lighting fluctuates.

Ariet's Algorithmic Innovation

Designed a four-stage pipeline: Hough Line fiducial alignment → 4-point homography transformation → dynamic subpixel bubble snapping (±15px tolerance) → adaptive brightness differential evaluation.

Python 3.12 OpenCV (cv2) NumPy Homography Matrix Tkinter Desktop UI Docker Compose
OLYMPIAD ANSWER SHEET
FORM #4920-KG • ACADEMIC EVALUATION
SERIES 02
Status: Click bubbles to edit marks, or test sheet tilt
Engineering Portfolio

Systems & Software Solutions

Applied software projects built to solve institutional, logistical, and commercial problems with modern web standards and asynchronous architectures.

Campus Infrastructure 2026

Ala-Too Booking AI — University Resource Management

Engineered a comprehensive campus facility reservation platform for Ala-Too International University. Replaces paper sign-ups and dispersed email threads with a real-time, conflict-free booking scheduler.

Architecture Highlights: Developed on React 19 and TanStack Start, deployed over Cloudflare Workers edge runtime. Utilizes optimistic state updates and Radix UI primitives for full keyboard and screen-reader accessibility.
React 19 TypeScript TanStack Start & Query Tailwind CSS v4 Cloudflare Workers Radix UI
Distributed Systems 2025

MagicWorlds — Bot Network Manager

High-throughput automation platform coordinating multi-bot tasks across hundreds of Telegram channels without hitting rate limits.

Dual-Tier Model Pattern: Django ORM acts as the database source of truth, while an asynchronous SQLAlchemy + Telethon pipeline executes high-speed worker queues.
Python 3.10+ Telethon MTProto SQLAlchemy Async Django + AdminLTE Docker Compose Redis & Postgres
Real-Time Logistics 2025

XAN-TAXI — Event-Driven Dispatch System

Automated taxi booking and dispatch bot system built on Aiogram 3. Handles driver queues, fare estimations, geolocation reverse-geocoding, and complete multi-language localization (i18n) for passengers and drivers.

Aiogram 3 Finite State Machines (FSM) Multi-language (i18n) PostgreSQL
Hardware Integration 2024 – 2025

Asia POS — Offline Retail Checkout System

Retail checkout desktop application interfacing directly with physical thermal receipt printers (XPrinter via ESC/POS protocol). Features offline data caching, inventory tracking, and automated sales auditing in Excel.

Python Desktop ESC/POS Thermal Protocol Hardware Driver Setup OpenPyXL Auditing
Preparedness & Capabilities

Technical & Academic Toolkit

A balance of rigorous algorithmic foundations, modern full-stack development, and international communication skills.

Algorithms & Vision

  • OpenCV (cv2) & NumPy Advanced
  • Homography & Perspective Warp Production
  • Hough Transform Line Detection Advanced
  • Subpixel Snapping & OMR Custom Algorithm
  • OCR (Optical Character Recognition) Proficient

Software Engineering

  • Python 3.12+ (Asyncio, Telethon) Advanced
  • React 19 & TypeScript Modern
  • TanStack Start & Query Modern
  • PostgreSQL & Redis Databases Proficient
  • Docker Compose & Linux Proficient

Languages & Global Mindset

  • English Academic / IELTS Prep
  • Kyrgyz Native
  • Russian Bilingual / Fluent
  • Turkish Fluent (Sapat / Ala-Too)
  • Acoustics & Sound Synthesis Cubase • FL Studio
Trajectory

Academic Journey & Milestones

A progression defined by curiosity, community service through code, and independent technical initiative.

2025 – PRESENT EdTech • Bishkek

Creator & Lead Engineer — AIM.EXAM OMR Pipeline

Initiated, architected, and field-tested the optical exam grading software now evaluated across thousands of school olympiad participants in Kyrgyzstan. Solved camera angle distortion and subpixel bubble detection without external dependencies or costly hardware.

2025 – 2026 Ala-Too International Uni

Full-Stack System Architect — Campus Space AI

Built modern resource booking platform on React 19, TypeScript, and Cloudflare Workers. Designed intuitive UI for scheduling university auditoriums, conference halls, and laboratory rooms.

2024 – 2025 Distributed Infrastructure

Backend Developer — Telegram High-Load Automations

Built scalable messaging infrastructure handling over 500,000 automated queries with Telethon, Django, and SQLAlchemy. Developed taxi dispatch systems with finite state machines and regional localization.

CREATIVE ROOTS Sound Production

Acoustic Signal Processing & Audio Synthesis

Multi-year exploration of digital signal processing in FL Studio and Cubase. Gained intuition for frequency spectrums, time alignment, and harmonic structure that directly reinforces mathematical computer vision work.

Direct Inquiries

Let’s Connect for Admissions & Collaboration

I welcome inquiries from university admissions officers, computer science professors, and research mentors. Based in Bishkek, Kyrgyzstan (UTC+6).