I'm a Computer Science student at San Jose State University (Class of 2026)
I'm passionate about AI, machine learning, and building tools that solve real-world problems
Currently a Break Through Tech AI Fellow and seeking internships & collaboration opportunities
Based in California | Open to remote and hybrid roles

I’m driven by the idea of using technology to create impact. Whether it’s building ML models that make predictions or developing backend services for real applications, I love bringing creative ideas to life. As a fellow in Break Through Tech AI, I've built ML pipelines, explored data science workflows, and contributed to team projects—all while growing my skills in Python, scikit-learn, and real-world engineering tools.

Python • Java • C • JavaScript • TypeScript • SQL • Shell

React • React Native • FastAPI • Node.js • Docker • Git • GitHub • PostgreSQL • AWS EC2 • Linux • VS Code • Jupyter Notebook • PySpark • Google Cloud Dataproc

scikit-learn • TensorFlow / Keras • pandas • NumPy • Feature Engineering • Hyperparameter Tuning • RAG • FAISS • Vector Databases • LLM Workflows • Prompt Engineering

REST APIs • SSE • MCP Servers • Security Tooling • Cloud Deployment

Prototype AI system simulating how security teams triage, prioritize, and respond to alerts using agent workflows and retrieval-augmented reasoning.

Tech Used: FastAPI, React, FAISS, OpenAI Embeddings, RAG, Docker

• Built an agentic AI workflow using:

  Perception → Reasoning → Action → Learning

• Implemented retrieval using FAISS vector search

• Added severity classification and alert routing

• Simulated phishing detection and SOC workflows

• Built historical incident retrieval for context-aware decisions

• Focused on explainability and practical security operations design

A machine learning system that predicts learner disengagement risk by combining EEG-derived brainwave features with self-reported survey data, developed in partnership with Chambers Capital Ventures as part of the Break Through Tech AI Studio
Tech Used: Python, scikit-learn, pandas, NumPy, Streamlit, dimensionality reduction, EEG feature engineering
Highlights:

• Built an end-to-end ML pipeline linking cognitive load and attention signals to engagement outcomes
• Achieved ~80%+ accuracy on held-out data using interpretable models (Logistic Regression, Random Forests)
• Identified early disengagement indicators to support adaptive learning interventions
• Designed a Streamlit dashboard to visualize engagement risk, EEG feature trends, and model explanations for non-technical stakeholders
• Conducted fairness checks across demographic subgroups to ensure transparent and equitable predictions

Academic Team Project — CS131 Processing Big Data

Large-scale fraud analytics system built on the PaySim dataset (~6M financial transactions).

Tech Used: PySpark, Google Cloud Dataproc, Pandas, UNIX Pipelines

• Fraud pattern analysis
• Data engineering workflows
• Feature engineering
• ML experimentation
• Distributed analytics processing

• Identified fraud transaction behaviors
• Generated transaction trend reports
• Built scalable analytics pipelines
• Explored financial risk analysis workflows

Academic Team Project — CS160 Software Engineering

Full-stack banking platform supporting transfers, account management, ATM search, recurring payments, and mobile access.

Tech Used: React, React Native, FastAPI, PostgreSQL, Docker, Plaid, AWS

• Mobile client development
• Transfer workflow implementation
• ATM locator integration
• Frontend components
• Testing and bug fixes

A full-stack, three-tier hospital management system designed to manage patients, doctors, appointments, and prescriptions
Tech Used: Java, JDBC, PostgreSQL, Apache Tomcat, Maven
Highlights: Built a database-driven healthcare system using JDBC with PostgreSQL and deployed it on Tomcat using a clean three-tier architecture

A multi-class classification model predicting a person’s workclass using U.S. Census data
Tech Used: Python, scikit-learn, pandas, KNN, Random Forest, GridSearchCV
Highlights: Streamlined data using a preprocessing pipeline and achieved over 80% accuracy

As a Fellow, I completed Machine Learning Foundations taught by Cornell faculty, where I:

• Built and tuned models (Random Forest, KNN, Gradient Boosted Trees)
• Developed supervised and unsupervised learning pipelines
• Gained practical exposure to real-world datasets and modeling workflows

Outside of coding, I mentor students in Python, Scratch, and Roblox Studio. I’m also active in the SJSU ACM Club and SWE++, helping lead outreach and tech initiatives. I enjoy building tech for social good and diving into conversations around AI ethics.

Email: ariansbahram@gmail.com
LinkedIn: linkedin.com/in/arian-bahram
GitHub: github.com/ariansbahram

I code best when listening to studio ghibli and collaborating with peers on meaningful tech!