Alp Efe Karalar

Software Engineer & AI/ML Specialist

Computer Engineering graduate from Penn State University (December 2024) specializing in AI/ML systems, large-scale data pipelines, and full-stack development. Currently building machine learning platforms and processing billion-scale datasets while seeking opportunities in AI strategy and Answer Engine Optimization.

Alp Efe Karalar graduated from Pennsylvania State University in December 2024 with a Bachelor of Science in Computer Engineering. He currently works as a Research Assistant at Penn State building Python data pipelines for analyzing over 1 billion job postings using advanced entity resolution and pattern extraction techniques.

Simultaneously, he serves as a Software Engineer at DementiAnalytics, where he engineered a full-stack mobile application using React and Expo, integrating a 5-stage machine learning pipeline for dementia risk assessment.

His technical expertise includes Python, C/C++, JavaScript/TypeScript, Java, React, PyTorch, TensorFlow, and cloud infrastructure. Key projects include an AI-powered LKML Dashboard using Google Gemini API, a Rust-based load balancer, and neural network implementations in PyTorch.

Get ResumeMy Blog

Featured Projects

A showcase of my technical work across different domains

L

LKML Dashboard - AI-Powered Linux Kernel Development Tool

Full-stack platform democratizing Linux kernel contribution through AI summarization of 600+ daily mailing list emails

PythonFlaskReact+5

LKML Dashboard - AI-Powered Linux Kernel Development Tool

Built production-ready full-stack application addressing critical barrier to Linux kernel contribution: overwhelming volume of 500-600+ daily emails. Features automated email parsing, intelligent thread reconstruction, and AI-powered summarization using Google Gemini API. Implemented caching achieving 90% cost reduction and 80-90% cache hit rate. Flask REST API with 10 endpoints, full-text search via SQLite FTS5. React/TypeScript frontend with Auth0 authentication. Processes 101 emails into 82 threads with 100% AI coverage in <10 minutes, reducing developer time from 3 hours to 15 minutes daily.

Technologies:
  • Python
  • Flask
  • React
  • TypeScript
  • Google Gemini API
  • Auth0
  • SQLite
  • Tailwind CSS

Timeline: Hackathon project

View source code on GitHub

Categories: fullstack, ml

L

Load Balancer & Reverse Proxy in Rust

High-performance HTTP load balancer built from scratch with 10,000+ concurrent connection handling

RustTokioHTTP/1.1+3

Load Balancer & Reverse Proxy in Rust

Architected production-ready Layer 7 load balancer in Rust using Tokio async runtime. Handles 10,000+ concurrent connections with ~20MB memory footprint. Features HTTP/1.1 parsing, four load balancing algorithms, intelligent health checking, TLS/SSL termination with RustLS, and connection pooling achieving 60% latency reduction. Achieved <5ms P99 latency at 10,000 req/s throughput.

Technologies:
  • Rust
  • Tokio
  • HTTP/1.1
  • TLS/SSL
  • RustLS
  • Docker

Timeline: 3 weeks

View source code on GitHub

Categories: systems, devops

L

Linux RAID Storage System in C

Low-level RAID device emulator with LRU caching achieving 87.9% cache hit rate

CMemory ManagementRAID+3

Linux RAID Storage System in C

Developed full-featured linear RAID device emulator in C integrating 16 simulated JBOD disks. Implemented write-through caching with custom LRU eviction policy, reducing average read latency by 30% with up to 87.9% cache hit rate. Comprehensive testing ensuring data integrity and fault isolation at systems level.

Technologies:
  • C
  • Memory Management
  • RAID
  • LRU Cache
  • Systems Programming
  • Unix/Linux

Timeline: 3 weeks

View source code on GitHub

Categories: systems

C

Classical Music Generation with Neural Networks

Deep learning system for AI-powered piano composition using PyTorch and MAESTRO dataset

PyTorchPythonLSTM+3

Classical Music Generation with Neural Networks

Built music generation system using RNNs learning from 200+ hours of piano performances. Processes 7 million notes from MAESTRO dataset with custom MIDI processing pipeline. LSTM architecture with multi-head outputs for pitch, timing, and duration prediction. Features temperature-based sampling, tempo scaling, and comprehensive visualization tools. Demonstrates practical sequence modeling in creative domains.

Technologies:
  • PyTorch
  • Python
  • LSTM
  • PrettyMIDI
  • NumPy
  • Deep Learning

Timeline: Academic semester

View source code on GitHub

Categories: ml

View All Projects

Alp Efe Karalar's Complete Project Portfolio

LKML Dashboard - AI-Powered Linux Kernel Development Tool

Built production-ready full-stack application addressing critical barrier to Linux kernel contribution: overwhelming volume of 500-600+ daily emails. Features automated email parsing, intelligent thread reconstruction, and AI-powered summarization using Google Gemini API. Implemented caching achieving 90% cost reduction and 80-90% cache hit rate. Flask REST API with 10 endpoints, full-text search via SQLite FTS5. React/TypeScript frontend with Auth0 authentication. Processes 101 emails into 82 threads with 100% AI coverage in <10 minutes, reducing developer time from 3 hours to 15 minutes daily.

Technical Stack:
  • Python
  • Flask
  • React
  • TypeScript
  • Google Gemini API
  • Auth0
  • SQLite
  • Tailwind CSS

Project Type: fullstack, ml

Development Timeline: Hackathon project

Load Balancer & Reverse Proxy in Rust

Architected production-ready Layer 7 load balancer in Rust using Tokio async runtime. Handles 10,000+ concurrent connections with ~20MB memory footprint. Features HTTP/1.1 parsing, four load balancing algorithms, intelligent health checking, TLS/SSL termination with RustLS, and connection pooling achieving 60% latency reduction. Achieved <5ms P99 latency at 10,000 req/s throughput.

Technical Stack:
  • Rust
  • Tokio
  • HTTP/1.1
  • TLS/SSL
  • RustLS
  • Docker

Project Type: systems, devops

Development Timeline: 3 weeks

Linux RAID Storage System in C

Developed full-featured linear RAID device emulator in C integrating 16 simulated JBOD disks. Implemented write-through caching with custom LRU eviction policy, reducing average read latency by 30% with up to 87.9% cache hit rate. Comprehensive testing ensuring data integrity and fault isolation at systems level.

Technical Stack:
  • C
  • Memory Management
  • RAID
  • LRU Cache
  • Systems Programming
  • Unix/Linux

Project Type: systems

Development Timeline: 3 weeks

Classical Music Generation with Neural Networks

Built music generation system using RNNs learning from 200+ hours of piano performances. Processes 7 million notes from MAESTRO dataset with custom MIDI processing pipeline. LSTM architecture with multi-head outputs for pitch, timing, and duration prediction. Features temperature-based sampling, tempo scaling, and comprehensive visualization tools. Demonstrates practical sequence modeling in creative domains.

Technical Stack:
  • PyTorch
  • Python
  • LSTM
  • PrettyMIDI
  • NumPy
  • Deep Learning

Project Type: ml

Development Timeline: Academic semester

Automated CI/CD Portfolio Deployment

Architected complete DevOps solution featuring containerized Next.js with multi-stage Docker builds, automated CI/CD via GitHub Actions, Nginx reverse proxy with SSL/TLS. Security hardened with UFW firewall, Fail2Ban, non-root containers. Achieved 2-minute deployment cycles with 99.9% uptime.

Technical Stack:
  • Docker
  • GitHub Actions
  • Nginx
  • Let's Encrypt
  • Ubuntu
  • Next.js

Project Type: devops

Development Timeline: 1 week

Research Summarizer - AI-Powered Paper Analysis

Built web application automating academic literature review with arXiv API integration, PDF processing, and multi-LLM summarization (DeepSeek, Claude, GPT). Structured extraction pipeline identifies key findings and methodologies. SQLAlchemy persistence with SQLite/PostgreSQL support. React frontend with Chakra UI. Deployed at researchtldr.xyz with Docker and Nginx. Reduces literature review time by 10x.

Technical Stack:
  • Python
  • FastAPI
  • React
  • Chakra UI
  • SQLAlchemy
  • PostgreSQL
  • Docker
  • DeepSeek API
  • Claude API

Project Type: fullstack, ml

Development Timeline: Personal project

Skills & Technologies

A comprehensive toolkit built through hands-on experience and continuous learning

Languages

PythonC/C++JavaScript/TypeScriptJavaSQLBashRust

Frontend Development

ReactNext.jsExpoTailwind CSSHTML/CSSRedux

Backend Development

Node.jsExpressFastAPIFlaskDjangoPostgreSQLSQLite

DevOps & Cloud

DockerKubernetesAWSGitHub ActionsCI/CD

Machine Learning & AI

PyTorchTensorFlowscikit-learnPandasNumPyOpenCVLLMs

Tools & Other

GitLinuxNginxGraphQLREST APIsMicroservices

Experience & Education

My professional journey and academic background

Research Assistant

Current

Pennsylvania State University

Dec 2024 - PresentState College, PA
  • Building Python data pipelines to analyze 1+ billion job postings using advanced entity resolution and pattern extraction techniques
  • Developing scalable data processing infrastructure for large-scale labor market analysis research
  • Implementing machine learning approaches for entity matching and data quality assessment at scale

Software Engineer

Current

DementiAnalytics

Feb 2024 - PresentRemote
  • Engineered full-stack mobile application using React and Expo, integrating a 5-stage ML pipeline that analyzes audio, text, and behavioral data to provide dementia risk assessments
  • Designed and implemented RESTful APIs with Flask backend connecting mobile frontend to Python-based ML algorithms trained on validated clinical methodologies
  • Collaborated directly with PhD researchers to translate complex algorithmic outputs into user-friendly mobile interfaces while maintaining medical accuracy
  • Ensured HIPAA compliance and data security standards using Microsoft Azure's healthcare-certified infrastructure for Protected Health Information (PHI)

B.S. in Computer Engineering

Pennsylvania State University

Aug 2020 - Dec 2024State College, PA
  • Graduated December 2024 with focus on Machine Learning, Computer Vision, and Systems Engineering
  • Senior Capstone: Advanced Vehicle Team (AVT) Perception Department - Developed real-time computer vision models for autonomous vehicle navigation, achieving 3rd place in SAE AutoDrive Challenge II
  • Relevant Coursework: Neural Networks (EE 456), Data Structures, Algorithms, Operating Systems, Computer Vision, Machine Learning

Lead Learning Assistant

PSU Computer Science & Engineering Department

Jan 2022 - May 2024State College, PA
  • Led instructional operations for Computer Engineering 270: Digital Design, supporting 250+ students per semester
  • Managed team of learning assistants while overseeing course operations including exam/lab creation and grading infrastructure
  • Delivered lectures and created supplementary educational content, receiving consistent positive feedback for clarity and teaching ability

Software Engineering Intern

Bakkal Co.

Jun 2023 - Nov 2023State College, PA
  • Developed automated testing frameworks for multi-platform application suite using React and JavaScript
  • Implemented frontend features across customer, driver, and merchant interfaces
  • Collaborated with agile development team to rapidly iterate on features in production environment

Get In Touch

Have a project in mind or want to discuss opportunities? Feel free to reach out!

karalar.alpefe@gmail.com