Henry Nawrocki

Designed and implemented a comprehensive AI-powered educational platform featuring multiple AI services and integrations.

Built conversational AI chatbots with context-aware responses using AWS Bedrock (Claude 3.5 Sonnet/Haiku) and OpenAI GPT models, implementing sophisticated prompt engineering and conversation management.

Developed AI exercise generation system with difficulty adaptation, content validation, and automated parsing of mathematical content including MathML processing.

Developed RAG integration architecture with AWS Bedrock Knowledge Base, implementing backend services for section-specific content retrieval, vector search querying, and relevance filtering.

Created voice AI features including speech-to-text transcription (OpenAI Whisper) and text-to-speech generation with audio file management and real-time voice chat interfaces.

Built AI content summarization service with multiple output formats and lesson content integration, plus AI-powered flashcard generation and tutoring hint systems.

Led a 4-person AI engineering team in designing and building a production-grade AI platform for a leading EdTech company — architected from scratch through deployment, shipped in weeks not months.

Architected and built a full-stack RAG (Retrieval Augmented Generation) system from the ground up: pgvector-powered semantic search across 64,000+ indexed content chunks, 14 specialized retrieval and generation tools, multi-model LLM orchestration layer, and a complete API service consumed by 6+ production applications.

Designed and implemented AI-powered educational features including conversational AI tutoring, automated exercise generation, intelligent content summarization, and an adaptive learning engine — integrated with AWS Bedrock (Claude models) and OpenAI APIs.

Engineered a content pipeline that migrated 18,000+ educational assets to S3 with automated metadata enrichment, enabling AI-driven content discovery and retrieval at scale.

Built and shipped a working adaptive practice system prototype that triggered an executive-level strategic review of the company's entire product direction.

Led technical strategy and architecture decisions as the primary AI technical leader, driving AI adoption across the full product suite while serving as technical liaison to C-suite executives, translating complex AI capabilities into business strategy and M&A positioning.

Technical Interviewer for Senior Developers specializing in Ruby on Rails, JavaScript, React, and Vue.js.

Conducted comprehensive technical assessments to evaluate candidates' proficiency in full-stack development technologies.

Mentored junior web developers, providing technical support through code reviews.

Proposed improvements to code organization to improve code quality and overall performance.

Provided advice and tips on maintaining motivation to maintain longevity in the program.

Co-founded and scaled a GPS fleet tracking business from zero to 60+ active enterprise clients across Peru, managing the full technical stack and operations.

Architected and built a real-time fleet management platform processing continuous GPS telemetry data from 60+ IoT devices — custom TCP/UDP parsers, PostgreSQL time-series storage, and live Google Maps visualization.

Designed server infrastructure on Digital Ocean and reduced operational costs from $10,000/month to ~$1,000/month (90% reduction) through strategic architecture decisions.

Managed B2B client relationships, hardware procurement, field installations, and a small operations team — full P&L responsibility for 7 years.