Ayse Sule Ekiz

Feb 2025 – Jul 2025

ForgeMind

Software Engineering Intern, Systems Performance & Data

• Profiled production Python services with py-spy, flame-graph analysis, and SQL execution-plan review to isolate CPU, memory, and query bottlenecks under live load.
• Applied index restructuring, join reordering, and CTE optimization, cutting dashboard render latency by 40% and reducing per-request CPU overhead across high-frequency reporting pipelines.
• Engineered resilient Python ETL pipelines ingesting REST APIs, web scrapers, and transactional databases into a centralized staging layer with schema validation and deduplication.
• Analyzed workload telemetry, feature-adoption funnels, and session engagement using window functions, CTEs, and cohort analysis; conclusions informed roadmap decisions across release cycles.

Sep 2024 – Present

North American University

IT & Systems Associate

• Shipped and maintained a high-reliability campus platform for NAU Athletics and student portal workflows, serving 1,000+ daily active users across students, staff, and administrators.
• Automated Python reporting workflows, eliminating 60% of manual processing overhead and giving staff faster access to live operational data.
• Instrumented A/B and usability testing cycles, connecting session data and conversion metrics to product decisions acted on directly by administrators.
• Built a reusable Figma design system and accessibility standards for campus platforms, reducing redundant design work and improving handoff quality.

Jun 2022 – May 2023

Koç University

AI Research Intern, Pipeline Optimization

• Diagnosed and optimized Python data pipelines for a 2D-to-3D deep learning reconstruction system, restructuring preprocessing and labeling stages to improve iteration speed.
• Reduced per-iteration compute overhead by about 30%, accelerating model evaluation cycles for graduate researchers.
• Built diagnostic JavaScript visualizations exposing model performance metrics and spatial transformation behavior for real-time inspection by non-technical stakeholders.

Aug 2024 – May 2027

North American University

B.S. Computer Science · GPA 3.95

• Leadership & Honors: President's Honor Roll · President, ACM-Women · GDSC Lead · President, Design Club · VP, Art Club

Aug 2018 – May 2023

Robert College

GPA 3.85 · Turkey's Premier American High School (est. 1863)

Spring 2026

Whisperwind Grove: Full-stack browser game platform

CS Senior Capstone · Full-stack game platform

• Built a TypeScript monorepo with React 19, Vite 7, Node.js, Express, Phaser 3, MongoDB, and Mongoose, hosting four playable browser games with shared player identity, achievements, and leaderboards.
• Integrated Gemini AI for gameplay evaluation and AI-assisted move selection, plus ElevenLabs-generated guardian voice clips; earned full marks for systems integration, AI implementation, and production-quality architecture.

May 2026

Framewise: AI video learning and dance practice workspace

HackHCC: Code Runners Hackathon · Best Use of AI · Best Use of MongoDB Atlas

• Built a connected React web app and Chrome MV3 extension sharing one Node.js/Express backend, MongoDB Atlas database, auth token, and Gemini AI pipeline for video analysis, chat, captions, notes, bookmarks, and quizzes.
• Trained a Keras BiLSTM classifier on MediaPipe pose landmarks, applied int8 quantization, and exported the model to TensorFlow.js for fully on-device browser inference.
• Added MoveNet-powered pose tracking for dance practice, with video/player overlays, webcam tracking, mirrored practice views, looped segments, speed controls, and session stats.

Apr 2026

Obi: Local AI desktop assistant

LA Hacks 2026 · Figma Make Challenge Winner

• Engineered a resilient, self-contained AI inference stack in 36 hours: locally-run Gemma LLM inside an Electron and Vite shell with a custom RAG retrieval pipeline.
• Removed external service dependencies from the critical path, achieving sub-second contextual responses with no cloud failure modes; judged production-ready by industry evaluators.

Feb 2026

Ghibli Guardians: AI-assisted personal safety companion

Rutgers University HackHERS Hackathon · Safety companion

• Built a React and Vite safety companion with Node.js and Express APIs for contextual risk assessment, trusted guardian contacts, location updates, and dynamically generated emergency messages with Google Maps links.
• Designed a calming character-based UX with chatbot-style safety guidance, copy/share actions, music, and ElevenLabs voice playback to make emergency preparation feel supportive rather than alarming.

Jan 2026

OpenStaxAlign: ML curriculum mapping tool

Rice Datathon 2026 · 2nd Place

• Architected an end-to-end ML classification pipeline, JSON ingestion, TF-IDF vectorization, Logistic Regression baseline, and Linear SVM, reaching 75% multi-class accuracy.
• Applied precision/recall diagnostics and confusion-matrix analysis to identify failure modes and guide iteration under hackathon constraints.
• Designed MongoDB schemas and analytics views that minimized query complexity and translated raw model outputs into a curriculum story that placed 2nd overall.

Dec 2025

Reverie: Mood-based creative spaces app

Full-stack web app · Personal project

• Created a React, Node.js, Express, and MongoDB application for mood-based creative spaces, supporting authentication, space CRUD, community discovery, saved collections, and timeline views.
• Designed the product in Figma first, then translated the visual system into a responsive interface for collecting reflections, images, color palettes, and emotional tags.

Sep 2024

NAU Athletics & Student Portal: Live campus platform

Full-stack platform · live in production

• Designed and shipped the NAU Athletics site and student portal serving 1,000+ daily active campus users, spanning UX research, design system, accessibility audit, developer handoff, and ongoing usability testing.