AI/MLBetta Health Oncology
International Medical & AI-Assisted Telemedicine Platform
Live client project at bettahealth.consulting • AI-assisted oncology intake & document classification • Secure RBAC & audit logging for medical compliance • End-to-end case management with procurement workflow
Challenge
Global oncology teams have no secure, compliant platform to manage international telemedicine cases. Manual case intake, scattered document handling, and unstructured patient data slow down critical workflows and create compliance risks.
Solution
Designed and shipped a full end-to-end oncology telemedicine platform: Case Intake → Stripe Payment Gating → Specialist Assignment → Interdisciplinary Tumor Board → Branded PDF Reporting. LLM-based extraction of structured oncology data (TNM staging, histology), automated document classification, granular RBAC (Patient/Doctor/CMO/Admin), patient consent gates, audit logs, and a procurement engine for oncology medications.
Next.jsTypeScriptSupabaseStripeOpenAI APIGitHub ActionsVercel
AI/MLEasyHR - AI Recruiting Platform
End-to-End Recruiting & Onboarding Automation
Automated end-to-end recruiting • AI-powered skill extraction • Explainable candidate scoring • Onboarding automation
Challenge
Traditional recruiting processes require manual CV screening, skill extraction, and candidate communication, leading to 40+ hours/week wasted by HR teams and qualified candidates being overlooked.
Solution
Built an AI-powered recruiting workflow that automates CV import, skill extraction, candidate ranking with explainable scoring, ATS pipeline management, and onboarding checklists. Integrated OpenAI/Gemini APIs for HR summaries and interview question generation.
Next.jsTypeScriptFastAPIPythonPostgreSQLRedisCeleryS3OpenAI APIGemini APIDocker
AI/MLRKA-AI - Knowledge Assistant
RAG System with RBAC & Source Citations
RBAC for secure knowledge access • Source citations & snippets • PII redaction & audit logs • Prompt injection defense
Challenge
Enterprise teams struggle to find information across scattered internal documents. Traditional search fails to provide contextual answers, and security concerns prevent AI adoption due to lack of access control and source transparency.
Solution
Developed a Retrieval-Augmented Generation (RAG) system with role-based access control (RBAC). Implements document pipeline (PDF/DOCX/HTML), chunking, embeddings, and vector search. Features audit logs, PII redaction, prompt injection defense, and always cites sources.
Next.jsTypeScriptSpring BootJavaPostgreSQLpgvectorRedisS3OpenAI APIGemini APIDocker
FullstackUrlaubPilot - Smart Holiday Planner
Trip Planning with AI Itineraries & Video Recaps
Price aggregation & scoring • AI itinerary generation • Vertical video recap studio • Async rendering pipeline
Challenge
Planning multi-destination trips is time-consuming, requiring hours of research across platforms. Travelers also lack tools to create shareable trip recaps for social media in vertical video formats (Reels/TikTok).
Solution
Created a full-stack travel planner that aggregates prices, scores optimal travel windows, generates AI-powered itineraries with recommendations and price alerts. Includes asynchronous video rendering pipeline using FFmpeg to create vertical recap videos from trip media.
Next.jsTypeScriptSpring BootJavaPostgreSQLRedisRabbitMQS3FFmpegDocker
