Scroll Code
A mobile code reader that turns a public GitHub repository into a commit pinned feed without executing repository code.
Sole engineer
Saber Garibi
AI systems software engineer
I build AI systems where permissions, evidence, and failure behavior matter, from governed pull request review and .NET services to an ESP32 camera that sends checked images to an iPhone.
Based in Chicago, Illinois
Selected work
Difficult problems with evidence attached
Three case studies that show how I work across product, system, and delivery boundaries.
iCan Eye
An accessible camera system that moves an image from an ESP32 device to an iPhone, selects an available vision path, and speaks a scene description.
Application and iCan Eye subsystem engineer
Evidence Bound Pull Request Review
A tested prototype that freezes pull request evidence at one commit, separates private analysis from public research, and returns structured findings for a person to judge.
Workflow designer and implementer
More work
Four more systems worth inspecting
Additional work spanning upstream security, fuzzing research, enterprise AI services, and recommendation systems.
NanoMQ Fuzzing Infrastructure
A merged upstream contribution that added a protocol aware libFuzzer path for the MQTT PUBLISH decoder and prepared a separate OSS Fuzz integration request.
Open source security contributor
Coverage Coach
An academic research prototype that turns missed MQTT source regions into diagnoses, typed interventions, replay checks, and auditable decisions.
Researcher and prototype engineer
Enterprise AI Service Boundaries
Sanitized professional work separating a growing AI application into authorized APIs, focused services, messages, data boundaries, and regression tested behavior.
AI software developer within a multidisciplinary team
Encore Recommendation Prototype
An internship prototype that transformed purchase records through a Python and SQL analysis pipeline into recommendation results, a working dashboard, and an executive presentation.
Data analyst intern and prototype developer
Engineering approach
Make the problem clear, keep the boundary small, and leave proof
Start with the operating constraint
Each case study names the condition that shaped the solution before it describes tools or implementation.
Explain the decision
Architecture and code are presented with the tradeoff they address, not as a list of technologies.
Keep claims inspectable
Public artifacts are linked directly. Private work uses sanitized diagrams and avoids invented performance or business numbers.
Experience
Production work across AI and full stack systems
Roles and responsibilities drawn from the verified candidate record.
August 2024 to present
AI Software Developer
CAI
Built group based access control across UserProfiles, replaced a stateful Model Context Protocol design with disposable clients, and fixed file propagation between project and chat workspaces.
September 2022 to July 2024
Full Stack Java Developer
CAI
Served as the sole developer with a project manager on a City of Chicago API integration that transformed records into the required database schema and closed a missing path for government users.
2023 to present
Independent Software Consultant
Appliance service client
Built performance focused Astro sites on Cloudflare with accessible contact flows, bot protection, consent controlled Google Places behavior, and automated browser checks.
June 2024 to August 2024
Data Analyst Intern
Encore Global
Queried purchase records from thousands of customers, moved the results into Python for segmentation, and presented an Apriori recommendation proof of concept and dashboard to leadership.
June 2022 to August 2022
Data Analyst Intern
CAI
Analyzed website behavior in Google Analytics, prepared reports and presentations, and compared analysis tools by capability, limitation, and use case.
September 2021 to June 2023
Work Study IT Associate
Wilbur Wright College
Diagnosed computer laboratory failures and prepared more than 100 machines through a repeatable system imaging workflow.
May 2025 to June 2025
Semiconductor Fabrication Trainee
University of Illinois Urbana Champaign
Carried a wafer through the complete four week fabrication process with a team while following cleanroom safety, process sequencing, and quality checks.
About
An engineer who works across the system
I build AI systems where permissions, evidence, and failure behavior matter, from governed pull request review and .NET services to an ESP32 camera that sends checked images to an iPhone.
My education combines computer engineering, cybersecurity, software, and data. My project work connects those disciplines to products people can use.
More about my backgroundProfiles