Jason Benjamin
Building automated data systems and custom digital products
New to Contra
Jason is ready for their next project!
Overview
Standardized testing and unit assessments often leave teachers with a mountain of raw test scores but very little actionable insight. "Benchmark Explorer" is a professional-grade student growth tracking and benchmarking system designed to eliminate the guesswork. It transforms raw academic data into clear, visual, and highly actionable dashboards that allow educators to identify learning gaps instantly and personalize instruction.
How It Works
The system acts as a centralized data repository where student diagnostic data, proficiency test scores, and daily assessment metrics are aggregated. The backend automatically processes these disparate data points, cross-references them against targeted learning benchmarks, and outputs clear visual trends that map out exactly where an individual student—or an entire cohort—is falling behind.
Technical Deep Dive & Guardrails
Relational Data Architecture: Built on a highly structured, relational Airtable backend optimized to handle multi-variable academic data without processing lag.
Automated Reporting Pipelines: Powered by complex Make.com (http://Make.com) logic pipelines that dynamically calculate student progress percentages, growth milestones, and historical target metrics upon new score ingestion.
Actionable Skill Mapping: Designed with conditional visualization layers that automatically flag specific sub-skills (e.g., phonics vs. reading comprehension) requiring immediate teacher intervention or curriculum adjustments.
0
15
Overview
Scheduling and resource allocation within busy educational environments is a notoriously complex problem. Administrators spend hours manually balancing room availability, teacher certifications, substitute mapping, and class rotations. EduPlanner replaces this chaotic process with an intelligent, automated scheduling engine that handles intricate constraints instantly.
How It Works
The system takes raw operational constraints—such as teacher availability, room capacities, curriculum tracks, and staff rotation rules—and processes them to generate optimal, conflict-free schedules. It actively flags potential issues and automatically suggests optimal substitute coverage or room re-mappings in real-time when changes occur.
Technical Deep Dive & Guardrails
Advanced AI Reasoning: Powered by a customized Google Gemini reasoning model that processes complex constraint matrices using recursive structural constraint re-weaving to find optimal scheduling paths.
Conflict Resolution Engine: Features a hard-coded algorithmic layer that enforces business rules with zero margins for error (e.g., ensuring a teacher is never double-booked across physical branches).
Real-Time Adaptability: Built to handle dynamic, live operational friction, such as last-minute sick leaves, by automatically recalculating and suggesting the lowest-impact substitute rotation.
0
18
Overview
Administrative overhead is one of the largest operational bottlenecks for modern academies and tutoring groups. Teachers often spend dozens of hours every month manually drafting student progress reports and cross-referencing records.
This project completely automates that entire operational loop, reducing teacher admin reporting time by an estimated 80% (saving over 15+ hours weekly per branch) while ensuring data remains clean and secure.
How It Works
Zero-Friction Intake: Teachers enter class review logs, grades, and behavioral notes directly through a clean, dynamic Fillout template at the end of a session.
Relational Routing: A specialized multi-step automation framework built in Make.com (http://Make.com) acts as the middleware. It instantly catches the form payload, deduplicates data, and runs strict validation checks.
AI Normalization: The data passes through a Google Gemini AI node, which automatically normalizes raw, unstructured teacher inputs into highly polished, concise bilingual summaries tailored for client-facing delivery.
CRM Sync: The fully compiled, structured data is mapped securely back into a relational Airtable cluster, dynamically linking records across student rosters and historical performance logs.
Technical Deep Dive & Guardrails
Relational Safeguards: Built custom data logic to retain historical student records seamlessly, even through unexpected mid-semester roster or teacher adjustments.
Row-Level Access Controls: Architected the database using strict row-level segregation controls, ensuring that administrative access is kept entirely separate from standard staff or public views.
Asynchronous Processing: Configured highly scalable, asynchronous webhook handlers capable of handling high-volume concurrent form submissions without a single point of failure or data loss.
0
20
Overview
Chekki AI is a mobile web application designed to eliminate the friction in "Mom-led English" homeschooling. Parents often lack the confidence to accurately explain English grammar and nuances to their young learners. Chekki solves this by acting as an on-demand, bilingual AI teaching assistant.
How It Works
Users simply snap a photo of any physical English worksheet or reading material. Using advanced image parsing, the app instantly overlays the digital answer key onto the screen. By tapping on any specific question, the user receives a customized, bilingual teaching script tailored to a young learner's level, complete with native-nuance explanations.
Technical Execution
AI Architecture: Powered by the Google Gemini API (1.5 & 2.0), utilizing custom JSON schema-structured prompts to guarantee highly accurate, bilingual parsing and translation.
Data Privacy Guardrails: Engineered with a strict "Zero-Memory Policy" to ensure 100% student data compliance—no personal data or homework images are stored post-processing.
Backend Infrastructure: Integrated with Firebase to handle secure, instantaneous payload routing and authentication.
The Impact
By automating the translation and pedagogical prep work, Chekki reduces lesson preparation time from 30 minutes to under 5 seconds per worksheet, empowering parents to teach with confidence and drastically lowering homework stress.
0
21