Processing...
NIX modernized the client’s eLearning platform and mobile app while extending its AWS-powered ecosystem with a Generative AI knowledge assistant that transformed institutional knowledge retrieval.
Healthcare, Education
Generative AI, Cross-platform Development
AWS, .NET, Angular, React Native, Python, Amazon Bedrock, Amazon Titan Embeddings
Our client, a leading provider of healthcare education software in the US market, understands the immense pressure medical schools face in today’s competitive landscape. Traditional methods often struggle to keep students up to date and may not adequately prepare them for complex exams or, more importantly, real work. This poses a significant challenge for graduates to be highly qualified professionals.
Head-on, our client tackles this challenge with innovative digital solutions. One of their products is a test simulation platform that assists medical students in their learning activities, providing them with valuable, exam-related content, analytics, and personalized feedback. The system also helps medical schools enhance their performance, giving them the means to improve educational outcomes and ultimately expand the workforce of healthcare professionals.
To meet growing user expectations for flexible, anytime-anywhere learning, the client needed to extend the platform with a mobile application while modernizing the existing system to improve scalability, performance, and speed of feature delivery through a robust, cloud-native AWS architecture. Internally, the company also needed to eliminate fragmented knowledge management processes by centralizing technical, operational, and training information into a searchable AI-powered system built on AWS, enabling secure, scalable, and intelligent access to institutional knowledge while reducing time spent searching for information and preserving critical expertise. The client went beyond their in-house resources and involved NIX in the project for its proven expertise in healthcare software and scalable AWS-based engineering solutions.
The client needed to modernize a mission-critical healthcare education platform while ensuring stable ongoing operations for medical schools and students already relying on it. Introducing new features for the platform and creating a new mobile application increased complexity, requiring tight synchronization with the existing system, a consistent cross-platform experience, and a development approach that would not slow delivery or compromise performance and reliability.
As the organization scaled, a second critical challenge emerged internally: fragmented and rapidly growing institutional knowledge. Years of technical decisions, bug fixes, onboarding materials, customer support insights, and training content were scattered across thousands of chat threads and recorded sessions, making information retrieval slow and inefficient. Engineers and support teams spent hours searching for answers, while valuable expertise was increasingly lost during employee turnover, creating operational risk and knowledge silos. This highlighted the need for a secure, scalable, and intelligent AWS-based solution that could consolidate and structure this “shadow knowledge” into a unified, continuously updated, and easily accessible source of truth.
NIX delivered two parallel engineering streams on AWS, addressing both platform modernization and internal efficiency within a unified, scalable architecture. The first stream modernized the eLearning platform and introduced a mobile application for seamless, on-the-go access to learning content and analytics. The second stream implemented a generative AI-powered internal knowledge retrieval system using Amazon Bedrock, transforming fragmented organizational knowledge into a centralized, intelligent search experience while reducing operational overhead and accelerating access to critical information.
In addition, NIX integrated a third-party adaptive learning service based on memory science principles such as spaced repetition and retrieval practice, enhancing learning effectiveness and enabling deeper behavioral data analysis to further improve personalization and user outcomes.
Our team upgraded the platform by migrating it to a newer .NET version for the backend and starting to use Angular for the frontend.
This dedicated and enhanced e-learning solution combines exam-style questions, comparative analytics, mnemonic videos, and individualized testing behaviors, helping institutions achieve better curricular and exam outcomes.
We developed a cross-platform mobile application that expands access to the examination system. Built with React Native, the app replicates core functionalities of the web platform, allowing students to practice exam tasks on their preferred devices, be it a smartphone—both iOS and Android,—desktop PC, or laptop. This way, they can engage in studying literally on the go.
The product has features that help students practice before high-stakes exams to be confident they’ll pass them easily. The functionality also ensures institutions can track, analyze, and improve learners’ performance outcomes.
The platform offers practice questions and exam-like tasks in various medical fields. Compiled by subject matter experts, they’re accompanied by simple and memorable bottom-line explanations.
This feature provides data-driven performance insights through national benchmarking that pinpoints curricular trends and shortfalls. With advanced reporting, the institution can guide faculty to improve learning for all, from entire cohorts to at-risk students, ultimately achieving program goals.
The system’s dashboard gives students access to informative metrics on their strengths and weaknesses in real time. This way, learners can see their category performance and benchmark themselves against peers nationwide.
The platform offers a test management service to constantly update questions, quizzes, and other tasks, ensuring they fully comply with the latest exam requirements.
The platform is equipped with a convenient search for questions and tests according to students’ requests, improving their learning experience.
To eliminate “search fatigue” inside the client’s organization, NIX designed and deployed a dual-pipeline retrieval-augmented generation (RAG) system on AWS. The solution ingests, structures, and serves answers from years of unstructured internal communications and recorded video content, making that knowledge instantly accessible to approximately 150 internal users without any manual searching.
Raw communication exports and meeting transcripts flow through a Python pipeline that masks personally identifiable information and restructures conversational threads. Amazon Nova Lite on Amazon Bedrock handles the high-volume preprocessing stage—a deliberate cost-conscious choice that keeps token spend predictable when processing large historical backlogs.
Rather than vectorizing raw messages as-is, every content chunk passes a “timelessness test” powered by chain-of-thought prompting via Anthropic Claude 3.5 Sonnet on Amazon Bedrock. This step distinguishes permanent technical knowledge (resolved bugs, configuration decisions, onboarding procedures) from ephemeral noise (greetings, logistics, off-topic chat). Only high-signal, “timeless” content advances to the knowledge base, ensuring retrieval quality remains high over time.
We designed a custom two-layer storage architecture on Amazon S3 to efficiently manage and search internal knowledge at scale. First, we generate lightweight semantic “fingerprints” of the content using Amazon Titan Text Embeddings v2 and store them separately from the full knowledge records. When a user makes a query, the system quickly searches these embeddings to identify the most relevant matches, and only then retrieves the complete, detailed information.
This approach avoids the need for a dedicated managed vector database, which would have been an unnecessary overhead for the client’s query volume. As a result, the solution keeps infrastructure costs very low while still delivering fast, sub-second response times and highly accurate search results.
We built a containerized Python backend using FastAPI, deployed on Amazon Elastic Container Service (ECS), which handles the core retrieval logic and coordinates all requests to Amazon Bedrock. This backend acts as the “brain” of the system, processing user queries and managing communication with AWS AI services. For the user-facing layer, we implemented a simple conversational interface using Streamlit, hosted on Amazon Lightsail, allowing internal teams to easily search and interact with the knowledge base. The architecture is intentionally decoupled: Lightsail provides stable and predictable monthly hosting costs for the UI, while Bedrock operates on a pay-as-you-go model for AI inference. This ensures the overall solution remains cost-efficient, scaling expenses only with actual usage.
Amazon Bedrock Guardrails enforce PII masking and content-safety policies at the generation layer, before any retrieved context reaches the model response. This ensures sensitive medical education data remains within the AWS trust boundary throughout the entire pipeline.
Before finalizing the architecture, we evaluated several alternative approaches. Enhancing traditional keyword search across communication platforms proved ineffective because it returned fragmented conversations without contextual synthesis.
Amazon Bedrock was ultimately selected for its serverless consumption model, multi-model ecosystem, and strong data privacy capabilities. The combination of Amazon Nova Lite, Anthropic Claude 3.5 Sonnet, and Amazon Titan Text Embeddings v2 enabled scalable processing, semantic reasoning, and efficient retrieval while maintaining strict compliance requirements for healthcare education environments.
Combined with Amazon S3 and Amazon ECS, the solution delivers a cost-efficient architecture that scales directly with actual usage.
Working side-by-side with the client’s team, we created a sophisticated e-learning ecosystem powered by advanced analytics, personalized learning experiences, and realistic exam simulation. The platform helps students improve performance through data-driven preparation, while giving institutions deeper visibility into educational outcomes and opportunities for continuous improvement. We continue to evolve the solution, helping medical schools better prepare highly qualified healthcare professionals.
Beyond the learning platform itself, NIX also transformed the client’s internal operations with a Generative AI–powered knowledge retrieval system. Tasks that previously required 2–3 hours of manually searching through chats and recorded sessions can now be completed in under one minute — reducing time-to-information by 99% for approximately 150 internal users. The solution also reduced dependency on senior engineers for onboarding and repetitive support requests, freeing up more than 1,500 engineering hours annually while creating a continuously updated, centralized source of institutional knowledge.
70+
Thousand Students
80+
Million Tests Taken per Year
+19.7%
Performance Improvement
99%
Reduction in Internal Knowledge Retrieval Time
1,500+
Engineering Hours Reclaimed Annually
AI Hazard Detection for Care Facilities: 98% Accuracy in Safety Threat Prediction
Healthcare
AI-Driven Application for Mental Health Support in the US
Smart Search Engine for Leading Automotive Marketplace
Retail and E-commerce
Automotive
Bridging OpenAI and Amazon Bedrock to Enhance LLM Evaluation Platform
Internet Services and Computer Software
AI Assistant: 70% Faster Support for Enterprise Employees
Social Networks and Communications
Starday Foods: Scaling to 100K Posts per Hour With AI
Food & Beverages
NIX delivers AI chatbot development services focused on security, scalability, and business impact—from strategy to production-ready AI solutions.
Harness the power of generative AI with our end-to-end services that include AI strategy, model training, deployment, and integration.
NIX is a software engineering company in the USA that offers enterprises digital transformation consulting services to embrace the future and growth.
Elevate your development capabilities with our turnkey team while you concentrate on growing your core business.
Schedule Meeting
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
