MedSchoolCoach
Boston, MA
Education Technology, Medical Admissions Consulting
Amazon Bedrock, Amazon S3, Amazon EC2, Amazon SageMaker, AWS Lambda, Amazon API Gateway
MedSchoolCoach empowers pre-med students with personalized one-on-one guidance from physician advisors with admissions committee experience. As the business scaled, the team needed a faster, more consistent way to extract key academic metrics from thousands of incoming student documents and turn them into actionable advising insights. Avahi delivered an AWS-based generative AI workflow that performs OCR on unstructured résumés and transcripts, normalizes GPA and MCAT information, and groups students into academic clusters based on historical success signals. This foundation enables advisors to spend more time on student engagement and deliver more specific, data-informed recommendations.
MedSchoolCoach is an education technology and medical admissions consulting organization that supports pre-med students through individualized advising, helping them build stronger applications using structured guidance, coaching, and admissions expertise.
MedSchoolCoach receives large volumes of résumés, transcripts, and academic documents every month. These inputs vary widely in structure and quality, including PDFs, scanned images, and screenshots, which makes automated extraction challenging.
The advising team was spending significant time manually extracting GPA, MCAT scores, coursework details, and academic patterns from these documents. This slowed student assessments, created review bottlenecks, and limited how consistently the team could personalize advising at scale.
Without automation, MedSchoolCoach would continue to face rising operational load as document volumes increased, which would limit growth and delay higher-value advising interactions.
MedSchoolCoach selected AWS to support a secure, scalable pipeline for ingesting and processing academic documents while using managed AWS services for model inference and analytics. AWS provided the foundation to store raw and processed documents, run clustering and evaluation workflows, and integrate foundation models through Amazon Bedrock.
MedSchoolCoach engaged Avahi because of Avahi’s experience delivering AWS-native generative AI systems that combine unstructured data extraction with repeatable analytical workflows. Avahi brought proven delivery patterns for integrating OCR with LLM reasoning, designing measurable validation steps, and packaging the solution behind an API-first backend that can evolve toward production-scale operations.
– AWS-hosted infrastructure on Amazon EC2 (including Elastic IP) to run the end-to-end pipeline
– Document ingestion and storage in Amazon S3
– OCR extraction pipeline for résumés and transcripts
– Automated parsing and normalization of GPA and MCAT metrics
– Unsupervised student clustering by academic similarity
– Amazon Bedrock integration for LLM-based interpretation and recommendation generation
– FastAPI backend endpoints for workflow execution and results delivery
– Technical documentation, demo handoff, and recommendations to scale to production
The engagement demonstrated that MedSchoolCoach can automate academic metric extraction and student segmentation workflows that were previously manual and time-intensive. By combining OCR with Bedrock-powered interpretation and clustering, the solution improves processing consistency and increases advisor capacity for higher-value student interactions.
Founder, Bravo Foxtrot
