AI Solutions Architect | Production AI Systems Developer | AI-Powered Engineering Evangelist
I'm an AI-native engineer who bridges enterprise finance expertise with cutting-edge AI development. Currently architecting production RAG systems and establishing AI-powered development methodologies that transform how teams work. My approach combines systematic AI collaboration with enterprise-grade quality standards, achieving team-level productivity as a solo developer.
Portfolio: lostmindai.com | Location: Sydney, Australia
Founder | AI Solutions Architect | Production Systems Developer
Production AI Systems:
- Architected LostmindAI Core - Production RAG system with semantic search and multi-LLM orchestration
- Implemented custom Cache-Augmented Generation (CAG) for efficient multi-file processing
- Built comprehensive AI platform serving real users with video analysis, chat, and content generation
- Deployed production systems with comprehensive monitoring and quality assurance
AI-Powered Development Methodology:
- Daily AI Collaboration: Expert user of Cursor, Windsurf, Continue, Cline, Claude Code
- Established systematic AI-human collaboration patterns maintaining quality standards
- Achieved productivity improvement while preserving code quality and innovation
- Created scalable development workflows that balance automation with human judgment
Technical Achievements:
- Multi-provider AI integration: OpenAI, Gemini 2.0, Claude, xAI with intelligent fallback
- Full-stack implementation: FastAPI backend + Next.js frontend + Vector databases
- Production deployment on Google Cloud Platform with Vertex AI integration
- Comprehensive testing framework ensuring enterprise-grade reliability
Mirvac Digital Team (2023-2024) - Solution Analyst & MRI Subject Matter Expert
- Led critical MRI to Yardi Voyager transition with zero downtime
- Developed Python automation reducing data migration time by 60%
- Promoted SQL optimisation tools across property management teams
Knight Frank Australia (2015-2022) - NSW Finance Manager
- Transformed financial operations through data-driven innovation
- Created SMAMS SQL Query Library (30+ queries) adopted as national standard
- Led culture change driving efficiency improvement across teams
- Mentored team of 5, establishing best practices for automation
My approach to transforming traditional development teams into AI-powered organisations:
Tool Adoption: Proven track record driving adoption of productivity tools across enterprise teams
Culture Change: Successfully led digital transformations in conservative financial organisations
Productivity Measurement: Documented efficiency gains through intelligent automation
Knowledge Transfer: Natural teacher and mentor for AI-enhanced development practices
Innovation + Quality: Systematic approach maintaining enterprise standards while embracing innovation
TECH_STACK = {
"AI_&_ML": [
"Production RAG Systems",
"LangChain Framework",
"Vector Databases (ChromaDB, Pinecone, FAISS, Vertex AI Vector Search)",
"Multi-LLM Orchestration (OpenAI, Gemini, Claude)",
"Prompt Engineering & Chain Design",
"Semantic Search & Embeddings"
],
"AI_Development_Tools": [
"Cursor (Expert)", "Windsurf", "Continue", "Cline",
"Claude Code", "GitHub Copilot",
"AI-Powered Development Workflows"
],
"Programming": [
"Python (Advanced)", "JavaScript/TypeScript",
"SQL Optimisation", "Modern Async Patterns",
"FastAPI", "Next.js", "React"
],
"Production_Systems": [
"Google Cloud Platform (Vertex AI)",
"Production Deployment & Monitoring",
"System Architecture & Design",
"API Development & Integration",
"Testing & Quality Assurance"
],
"Data_&_Analytics": [
"Vector Database Design",
"SQL Performance Optimisation",
"Data Pipeline Architecture",
"Business Intelligence"
],
"Enterprise_Experience": [
"MRI (Subject Matter Expert)",
"Yardi Voyager", "Oracle JDE",
"Financial Systems Integration",
"Change Management"
]
}
Repository: gemini-mcp-server
Problem: Integrating cloud services into Claude Code without complex local configuration
Approach: Lightweight wrapper pattern separating MCP interface from business logic
Outcome: Reusable template for API/service integration
Technologies: TypeScript, MCP Protocol, Cloud Run, Google Gemini API
Repository: gemini-cloud-service-v2 (Private)
Problem: Processing 50+ file types with AI while maintaining performance and scalability
Approach: Cloud-native FastAPI service with optimised Gemini integration
Outcome: Handles massive files that exceed local processing capabilities
Technologies: Python, FastAPI, Google Cloud Run, Gemini 2.0 Flash
Repository: LostMindAI-Core
Live at: lostmindai.com
- Production-grade RAG implementation with semantic search
- Multi-modal content processing (text, video, documents)
- Custom CAG (Cache-Augmented Generation) for performance
- Full-stack: FastAPI + Next.js + Vector databases
- Comprehensive testing and monitoring
Repository: LostMindAI-Enterprise-Platform
Problem: Enterprise-grade AI content analysis with cost optimisation and monitoring
Approach: Advanced platform with thinking budgets, caching, real-time monitoring
Outcome: Production platform with intelligent caching and advanced RAG capabilities
Technologies: Gemini 2.5, Advanced RAG, Cost Optimisation, Real-time Monitoring
Repository: What-is-AI-Agents
Problem: Comprehensive knowledge base for AI agents bridging theory and implementation
Approach: Educational content and documentation covering conceptual to practical implementation
Outcome: Knowledge sharing platform for AI agents by LostMind AI
Technologies: Documentation, Educational Content, AI Agent Concepts
Repository: sumit-mondal-cv
Problem: Professional CV and resume system with proper documentation
Approach: LaTeX-based system with version control and professional formatting
Outcome: Professional CV system with comprehensive documentation
Technologies: LaTeX, Documentation, Version Control
- Bachelor of Commerce - Accounting (Macquarie University)
- Self-Directed AI/ML Specialisation (2024-Present)
- Production RAG system architecture
- Multi-LLM orchestration patterns
- AI-powered development practices
- 15+ years Enterprise Finance & Systems
- Daily learner in emerging AI technologies
I'm passionate about:
- Building production AI systems that solve real problems
- Transforming how teams work with AI-powered development
- Measuring and improving developer productivity
- Creating scalable, maintainable AI architectures
- Teaching and promoting AI development practices
Open to: AI architecture roles, developer productivity positions, technical leadership opportunities
- Portfolio: lostmindai.com
- LinkedIn: Connect with me
- Email: Contact me
"Not just writing code with AI - architecting the future of how we build software"