Muhammed Cengiz
RESUME
AI/ML
Web

Feather — AI Market Insights Platform (Capstone, Team)

Team capstone building an AI-assisted next-day stock prediction workflow with clear evaluation, safeguards, and a web UI. Capstone prototype graded 94/100.

Live demo← Back to library

At a glance

Role

Team

Type

Capstone

Status

Graded 94 / 100

Year

2025

Stack

6 technologies

Links

Live · Case study

Media

Achievements

3 of 4 unlocked

  • Evaluation Owner

    Built the evaluation harness — baselines, metrics, the whole thing.

  • Contract Driven

    Defined the API between model and frontend so neither team blocked the other.

  • Team Player

    Shipped as part of a 4-person team — code reviews, sprint planning, all of it.

  • Hidden achievement

    Keep playing this project to find out.

Overview

Problem

Users wanted a simple UI for exploring signals and predictions with transparent evaluation and limits on overconfident outputs.

My role

Team contributor: owned ML evaluation + API contract; implemented feature work end‑to‑end and documented trade‑offs.

Stack

React + TypeScript (Vite) • FastAPI (Python) • SQLAlchemy • Docker • ML baselines (SVR/RF) • Vercel (frontend)

Highlights

  • Owned the evaluation approach (baselines + metrics) and translated results into clear UI-friendly outputs.
  • Defined and implemented the API contract between model code and the web UI for predictable integrations.
  • Capstone prototype graded 94/100.

Case study

Feather — AI Market Insights Platform (Capstone, Team)

Live demo: https://www.feathertrade.org/

Summary

Full-stack capstone app for next-day market insights: baseline ML predictions plus optional sentiment features, packaged into a UI that’s easy to explore. The codebase now lives in a private repository as the project continues past the capstone.

My contribution

  • Owned evaluation approach (baselines + metrics) and translated results into clear UI-friendly outputs
  • Defined the API contract between the backend and frontend
  • Built features end-to-end (data → API → UI) and documented trade-offs for the team

Tech snapshot

  • Frontend: React + TypeScript (Vite)
  • Backend: FastAPI (Python) + SQLAlchemy
  • Docker for local/dev setup