From grading papers tounderstanding students
Every answer sheet we evaluate teaches the AI what good reasoning looks like. Grading is the data engine; understanding student potential is the destination.
Subjective evaluation is the bottleneck
Consider the math of a typical school or test-prep institution. An academy with 840 students and 20 teachers conducting 4 descriptive mock exams per year generates 3,360 answer booklets per exam cycle.
Each descriptive paper takes 8 to 10 minutes to grade properly. This means teachers spend 3 weeks of grading admin per exam cycle—equating to 140 to 300+ hours spent grading per teacher every year.
As a result, 32% of teachers have considered quitting due to the administrative workload, while students wait weeks for feedback that is already outdated. They don't just need scores; they need to know why they are wrong.
Formative assessment makes AI grading inevitable
Modern educational standards globally—from ESSA and Common Core in the US, to Ofsted in the UK, and NEP 2020 in India—mandate a structural shift from high-stakes summative exams to regular, formative, competency-based continuous assessments (IB, AP, IGCSE, CBSE).
Moving from a few major exams a year to continuous formative evaluations represents a 10x increase in grading volume. With global teacher workloads at an all-time high, educators carry this administrative load alone.
The math simply does not work without automation. AI-powered subjective grading assistants act as a scaling layer, enabling schools to run continuous evaluations without burning out their teaching staff.
Empowering teachers, unlocking potential
Ayush Poojary
Founder, GradeSense
GradeSense was born out of a simple observation: the grading system shouldn't just rank students; it should help them learn. Having completed rigorous academic tracks across CA, CFA, and MBA programs, founder Ayush Poojary witnessed firsthand how schools and teachers are buried under grading administrative work, leaving little capacity for personal mentorship.
We believe that when you free teachers from hundreds of hours of manual grading, you return that time to the classroom. GradeSense uses logic-aware AI to evaluate descriptive answers (explore our features), allowing institutions to provide high-quality, continuous feedback to every student at a fraction of the time and cost (view our early access pricing).
Our mission is to build the feedback infrastructure that democratizes education—making personalized, step-by-step guidance accessible to every learner so they can achieve their true potential.
GradeSense is incubated at Scaler and supported by Google for Startups, Microsoft for Startups, and AWS Activate.
Our roadmap to understanding students
Evaluation Engine
AI grades handwritten sheets using teacher rubrics and model keys. Currently running pilots with 100% manual check option.
Semantic Analysis
AI learns subject syllabus deeply enough to evaluate reasoning steps and award marks without requiring strict model keys.
Personal Companion
AI tracks each student's strengths, gaps, and pace. Moves beyond scoring to become a custom learning companion.
“Every student, regardless of their starting point, deserves to know not just what their score is, but how they think. We are building the feedback infrastructure to make personalized learning guidance a basic right, not a privilege.”
The Long-term Vision of GradeSense
Every student deserves a grade that understands them
Join us in making instant, logic-aware evaluation the global standard. Free during pilot launch.