Arda Edil

Student · Builder · Founder

Arda Edil

I build software that makes AI systems more useful in the real world.

I'm a Computer Science student at the University of Michigan working across AI, full-stack engineering, and applied ML. Right now, I'm building AgentSEO and experimenting with everything from AI-agent reliability to reinforcement learning and computer vision.

✦AI / ML⧉Full-Stack↗Founder @ AgentSEOMMichigan CSE
Explore my workResume ↗

What I'm doing right now

NOW / AUGUST 2026

Building

AgentSEO

AI-agent/API reliability tooling

Engineering

CaseForge @ BGTS

AI-powered software engineering

Exploring

AI Agents, RL & Computer Vision

Applied AI experimentation

Studying

Computer Science @ Michigan

Engineering + Business Minor

PORTFOLIO AI

Ask my portfolio anything.

Don't want to scroll through everything? Ask my AI about my experience, projects, research, or technical background.

Grounded in Arda's portfolio content not a general-purpose chatbot

Experience

Founder - AgentSEO
2026 - Presentview repo
  • Building a startup product that measures how reliably AI agents operate OpenAPI-based interfaces before touching production systems.
  • Designed a benchmark workflow for importing OpenAPI specs, normalizing tools, generating tasks, running provider adapters, and comparing deterministic outcomes.
  • Implemented a FastAPI, PostgreSQL, Next.js, and Tailwind stack with sandbox resets, visible traces, failure taxonomy, tests, Docker Compose, and keyless CI.
Software Engineering Intern - BGTS
Jun 2026 - Aug 2026view repo
  • Owned the CaseForge librarian module, which maps RFP requirements to the past engagements that best prove BGTS capabilities.
  • Built ranked matching with rationale support so proposal teams could find relevant evidence under deadline pressure.
  • Defined JSON contracts and implemented validation, error handling, and unit tests across team-owned pipeline components.
Project Manager & Software Developer - Menlo Innovations
Sep 2025 - May 2026
  • Led development of a timesheet data system from stakeholder requirements through technical delivery.
  • Translated requirements into SQLite and SQLAlchemy schemas for reliable operational data storage.
  • Coordinated Python data-import work across the project team and kept implementation aligned with client needs.
Supply Chain Analyst Intern - Setas Masterbatch
Jun 2025 - Aug 2025
  • Replaced a k-means demand-segmentation baseline with agglomerative-clustering models in Python and R.
  • Improved forecast accuracy by 13% and increased on-time material availability by 8 percentage points.
  • Reduced stockouts by 20% and enabled purchasing decisions 10 days earlier.

Projects

Founder & Full-Stack Engineer - AgentSEO
2026Startup + AI Agentsview project
  • Built a working research MVP for measuring how interface design affects AI-agent reliability on OpenAPI-based APIs.
  • Created a benchmark system with normalized tools, generated task catalogs, resettable sandboxes, deterministic assertions, failure taxonomy, and cross-model reports.
  • Shipped a FastAPI/PostgreSQL backend and Next.js workflow UI with OpenAI, Anthropic, Gemini, and bounded mock-agent adapters.
Machine Learning Engineer - MarioMind
2026Deep RLview project
  • Built a reproducible reinforcement-learning lab for controlled World 1-1 experiments with DQN and Double DQN agents.
  • Implemented configurable reward shaping, frame stacking, exploration schedules, replay buffers, and action spaces.
  • Added random, always-right, and reflex baselines plus experiment plots and failure summaries for meaningful comparisons.
Full-Stack ML Engineer - VARLens AI
2026Video AIview project
  • Built an educational soccer clip analysis MVP that estimates foul sanctions and action types with uncertainty-aware explanations.
  • Connected a Next.js interface to a FastAPI service with shared JSON and TypeScript/Python contracts.
  • Added VideoMAE checkpoint serving, transient upload handling, automated tests, and deterministic fallbacks for reliable development.
AI Engineer - AI Stock Analyzer (RAG + Retrieval)
2024AI / RAG
  • Built a retrieval-augmented generation pipeline using LangChain and FAISS for grounded financial Q&A.
  • Used ReAct-style prompting and retrieval filtering to reduce hallucinations and improve response quality.
  • Improved answer quality by 18% in evaluation and packaged the results into a usable workflow.
Research Assistant - OxPal - Biofeedback Research
2024Published Researchview project
  • Designed a child-friendly biofeedback system using PPG-based HRV sensing, Bluetooth communication, and guided breathing.
  • Contributed to the system architecture and interactive interface with a multidisciplinary research team.
  • Presented the work at HCI International 2024 and contributed to a Springer CCIS publication.
Software Engineer - Lore
2025Mobile + Backendview project
  • Worked on an Innovation for Impact mobile app for University of Michigan student community building.
  • Built React Native and Expo user-facing flows, navigation patterns, onboarding, and reusable UI components.
  • Integrated frontend features with Django REST Framework backend APIs and typed OpenAPI-generated contracts.
Full-Stack Engineer - AI-Powered Portfolio Website
2026Full-Stack + AI
  • Built a Next.js portfolio with an AI chat panel powered by a custom API route.
  • Implemented retrieval over portfolio content to provide grounded, first-person answers.
  • Designed a founder-first opening, keyboard command palette, grounded AI interface, and responsive project cards.

Technical Coursework

EECS 492 — Introduction to Artificial Intelligence

Search, planning, probabilistic reasoning, and machine learning foundations for building intelligent systems.

EECS 376 — Foundations of Computer Science

Theory of computation, automata, computability, and complexity—formal reasoning about what problems are solvable.

EECS 370 — Computer Organization

How software maps to hardware—assembly, memory hierarchy, processors, and performance-aware systems programming.

EECS 281 — Data Structures & Algorithms

Algorithm design and analysis using trees, graphs, hash tables, and priority queues, with a focus on performance tradeoffs.

EECS 280 — Programming & Data Structures

Object-oriented programming in C++, data abstraction, memory management, and building medium-scale software systems.

EECS 203 — Discrete Mathematics

Logic, proofs, sets, graphs, and combinatorics—foundations for algorithms, correctness, and complexity reasoning.

EECS 201 — Computer Science Pragmatics

Unix, scripting, build systems, version control, debugging, and tooling used in real-world software development.

Technical Toolkit

Skills

Languages, frameworks, AI tools, and delivery practices
Python
Backend / Scripting
TypeScript
Full-Stack Apps
JavaScript
Web Development
React
Frontend UI
Next.js
Portfolio / Apps
Tailwind CSS
Interface Styling
Node.js
Backend Services
FastAPI
Python APIs
SQLAlchemy
Data Modeling
SQLite
Databases
PyTorch
Machine Learning
LangChain
AI Workflows
FS
FAISS
Vector Search
Java
Programming
C/C++
Systems Foundations
R
Data Analysis
Git & GitHub
Version Control
OK
Testing
Reliable Delivery