Selected engineering work

My Journey.

Exploring ideas, building projects, and learning along the way.

Total projects

8 artifacts

Timespan

2025 — Present

Current focus

Secure systems & automation

8 projects

2026

Current focus

An end-to-end multi-tenant lending platform covering lenders, borrowers, collectors, loans, repayments, QR workflows, reporting, and SMS notifications.

MORTGAGE / 01Project artifact

Secure lending operations

PORTFOLIO

Key contribution

Designed tenant isolation and authorization, secure session handling, transactional payment workflows, duplicate-request protection, and automated quality gates.

Client use: ~1 year
Architecture: Multi-tenant
Delivery: CI tested

Next.jsTypeScriptAppwriteVitestPlaywrightGitHub Actions

GitHub Live demo LinkedIn

An asynchronous Telegram assistant for manual and AI-assisted email drafting, approval workflows, job-application generation, and secure Gmail delivery.

AI MAIL / 02Project artifact

Human-in-the-loop automation

PORTFOLIO

Key contribution

Built a provider-independent LLM layer, grounded structured generation, Google OAuth with PKCE, encrypted credential storage, explicit state machines, and duplicate-send protection.

Tests: 123
Delivery: Multi-user
Security: OAuth + PKCE

PythonOpenAI APITelegramGmail APIOAuth 2.0SQLitepytest

GitHub

A multi-role credit management platform developed as a university team project, with lender-facing workflows, authentication, and SMS notifications.

CREDIT / 03Project artifact

Collaborative credit workflows

PORTFOLIO

Key contribution

Focused on the lender web application, shared NestJS REST API integration, Firebase data flows, authentication coordination, and cloud deployment.

Grade: A+
Roles: Multi-role
Backend: Cloud Run

ReactTypeScriptNestJSFirebaseGoogle Cloud RunVercel

GitHub

An asynchronous Python utility for processing queued Telegram message links, retrieving accessible media, organizing files, and reporting transfer progress.

TELEGRAM / 04Project artifact

Queued media processing

PORTFOLIO

Key contribution

Structured Telegram API access with environment-based configuration and reusable asynchronous queue-processing logic.

Runtime: Async
Processing: Queued
Config: Env based

PythonTelethonasyncioTelegram API

GitHub

A mobile-first storefront with product and category pages, search, image galleries, administrative workflows, and registration-free WhatsApp ordering.

COMMERCE / 05Project artifact

Conversation-led commerce

PORTFOLIO

Key contribution

Built protected product, order, and store-management workflows with Appwrite real-time data, multi-image storage, Zod validation, and optimized catalog queries.

Ordering: WhatsApp
Experience: Mobile-first
Validation: Zod

Next.jsReactTypeScriptTailwind CSSshadcn/uiAppwriteZod

GitHub

2025

A responsive course catalog with detailed learning pages, preview videos, pricing and level information, plus a complete course administration dashboard.

COURSE / 06Project artifact

Structured learning delivery

PORTFOLIO

Key contribution

Developed course CRUD workflows and integrated Appwrite Database and Storage for course data, uploaded media, and external video URLs.

Content: Video ready
Admin: Full CRUD
Hosting: Vercel

Next.jsReactTypeScriptTailwind CSSAppwriteReact PlayerVercel

GitHub Live demo

A 12-servo quadruped robot featuring multi-gait locomotion, web and mobile control, obstacle detection, calibration tools, and interactive behaviors.

ROBOT / 07Project artifact

Gesture-controlled robotics

PORTFOLIO

Key contribution

Led the ESP32-CAM gesture-recognition module, designed 3D-printable components, configured the Raspberry Pi environment, and contributed to MPU-6050 PID stabilization.

Actuation: 12 servos
Control: Multi-gait
Stability: PID

PythonRaspberry PiESP32-CAMOpenCVMediaPipeMPU-6050WebSockets

GitHub LinkedIn

A real-time computer-vision system that interprets hand gestures from an ESP32-CAM video stream for low-cost gesture-based interaction.

VISION / 08Project artifact

Real-time gesture tracking

PORTFOLIO

Key contribution

Combined embedded video streaming with MediaPipe landmark detection and OpenCV frame processing to recognize predefined gestures from live input.

Input: Live video
Vision: MediaPipe
Device: ESP32-CAM

PythonOpenCVMediaPipeESP32-CAMArduinoC++

GitHub