Contact Information:

Building enterprise-grade systems with AI at the core — multi-provider LLM pipelines, ML models, agentic orchestration, and full-stack architecture

Full Stack Enterprise Architect with 15+ years designing and delivering scalable, application-heavy enterprise solutions. Deep expertise in .NET Core, Python, React, Angular, and multi-cloud platforms (Azure + AWS). Builds production-grade Generative AI and Agentic AI systems — multi-provider LLM pipelines (Claude, Gemini, OpenAI), LightGBM ML models, record linkage for entity deduplication, and autonomous orchestration pipelines. Consistently delivers measurable results: 60% performance improvements, 25% infrastructure cost reductions, 99.9% uptime for platforms handling 100K+ daily requests.

• Generative AI & ML Systems: Multi-provider LLM integration (Claude/Gemini/OpenAI), LightGBM, record linkage, agentic pipeline orchestration, prompt engineering
• Enterprise Architecture: Designing scalable, application-heavy systems with microservices, event-driven patterns, and fault tolerance
• Full Stack Development: .NET Core, Python/FastAPI, React, Angular, Vue.js — backend to frontend
• Multi-Cloud: Azure (primary) — App Services, Service Bus, Container Instances, Bicep IaC; AWS — EC2, Lambda, IAM
• Data Engineering: SQL Server, MongoDB, Redis, MySQL, PostgreSQL — optimisation, indexing, caching strategies
• Team Leadership: Successfully led development teams of 5–15 developers across multiple enterprise deliveries

Back to Top

Back to Top

Back to Top

Enterprise booking API serving over 100,000 daily requests

This comprehensive travel booking system powers multiple production websites including www.hop-on-hop-off-bus.com. Built with .NET 6, it integrates with over 20 major travel suppliers including Tiqets, Bokun, BigBus, HotelBeds, and GetYourGuide.

Key Features:

• Real-time booking and availability checking
• Multi-language support for international customers
• Automated voucher generation and delivery
• Background processing with Hangfire
• Comprehensive caching layer using Redis
• Clean architecture with extensive testing

Event-Driven Architecture:

• Azure Service Bus for reliable message queuing and pub/sub messaging
• Azure Logic Apps for workflow automation and system integrations
• Asynchronous booking confirmations and notifications
• Decoupled microservices communication

Technical Implementation:

• .NET Core backend
• RESTful APIs with Swagger documentation
• JWT authentication and rate limiting
• SQL Server for data persistence
• Azure cloud deployment
• Entity Framework Core for data access

Advanced hotel pricing optimization and automated rebooking system

This intelligent system processes over 10,000 hotel reservations daily, continuously monitoring prices across multiple suppliers to automatically rebook customers at lower rates, achieving 15-30% cost savings.

Key Features:

• Multi-supplier price comparison algorithms
• Automated rebooking at lower costs
• Machine learning-based optimization
• Real-time price monitoring
• Dynamic pricing algorithms

Technical Implementation:

• .NET Core backend with Azure Batch processing
• SQL Server and MongoDB for data storage
• Hangfire for background job processing
• Redis for high-performance caching
• Integration with multiple hotel booking APIs
• Comprehensive notification systems

Comprehensive multi-tenant property and facility management platform

An enterprise-grade system supporting over 50 properties and 1000+ facilities with complete management capabilities.

Key Features:

• Multi-tenant architecture for scalability
• Gate management and access control
• Inventory tracking and management
• Advanced reporting and analytics
• Responsive frontend interface

Technical Implementation:

• .NET Core backend with SQL Server
• Angular 15+ frontend framework
• Docker containerization
• Azure DevOps for CI/CD
• Multiple deployment environments (Belt, Karamat, Leather)

Luxury resort and hospitality website with modern design

A high-performance resort website featuring modern design principles and achieving 95+ Lighthouse performance scores.

Key Features:

• Responsive design for all devices
• Interactive image galleries
• Amenities showcase and booking
• Contact management system
• Smooth animations and transitions

Technical Implementation:

• Next.js 13+ with React 18
• TypeScript for type safety
• Tailwind CSS for styling
• GSAP for animations
• Static site generation for performance
• Render.com deployment with CDN

Comprehensive railway operations and passenger information system

A complete railway management solution handling operations across multiple stations with real-time passenger information and multi-language announcements.

Key Features:

• Train scheduling and management
• Station operations control
• Passenger information system (IPIS)
• Multi-language voice announcements
• Real-time status updates

Technical Implementation:

• .NET Framework with Windows Forms
• SQLite and Access Database integration
• Web API for external integrations
• Audio system integration
• Desktop applications for station management
• 100+ pre-recorded announcements in multiple languages

Enterprise multi-marketplace synchronization and logistics platform

• Backend: Node.js, Express.js, MySQL, JWT authentication, comprehensive Swagger API documentation
• Frontend: React 18, Vite, Tailwind CSS, responsive design with real-time updates
• Architecture: Microservices architecture with separate frontend/backend, automated cron jobs, error management

Event-Driven Architecture:

• Apache Kafka for high-throughput real-time event streaming
• AWS SQS for reliable message queuing across services
• Event-driven inventory sync and order processing
• Real-time marketplace updates with pub/sub patterns

Marketplace & Shipping API Integrations:

E-commerce Platforms:

• Amazon SP-API: Product sync, inventory management, order processing, price updates
• Shopify REST Admin API: Product variants, location sync, stock updates, order fulfillment
• WooCommerce REST API: Product catalog sync, order management, tracking integration
• Magento REST API: Product synchronization, stock management, order processing
• PrestaShop Webservice API: Multi-store product sync, inventory updates, order tracking
• eBay Developer APIs: Listing management, inventory sync, order fulfillment
• Mirakl Marketplace APIs: Multi-vendor product sync, carrier management, order tracking

Shipping & Logistics APIs:

• BlueDart Integration API: Express shipping, real-time tracking, delivery management
• DTDC Express API: Courier services, shipment tracking, delivery notifications

Core Functionalities:

• Product Synchronization: Bi-directional sync across all marketplaces with variant support
• Inventory Management: Real-time stock updates, low-stock alerts, automated reordering
• Order Processing: Centralized order management, status updates, automated fulfillment
• Price Management: Dynamic pricing, bulk price updates, marketplace-specific pricing
• Tracking Integration: Multi-carrier tracking, automated status updates, customer notifications
• Error Management: Comprehensive error logging, retry mechanisms, alert systems
• Cron Automation: Scheduled sync jobs, automated data processing, background tasks

Scale & Performance:

• Multi-tenant architecture supporting 100+ clients simultaneously
• Real-time synchronization across 8+ major marketplaces
• Automated processing of 10,000+ products and orders daily
• Comprehensive API documentation with 150+ endpoints
• Advanced error handling with retry logic and alert management

Modern web application with Vue.js ecosystem

• Tech Stack: Vue.js 3, Nuxt.js 3, TypeScript, Tailwind CSS, Vite
• Features: Server-side rendering, responsive design, modern UI/UX, PWA capabilities
• Deployment: Azure Static Web Apps with automated CI/CD pipelines
• Performance: Optimized build process, lazy loading, and efficient asset management

Infrastructure as Code for multiple production environments

• Tech Stack: Azure DevOps, YAML pipelines, ARM templates, PowerShell scripts
• Environments: Multiple PMS environments (Belt, Karamat, Leather) with environment-specific configurations
• Features: Automated deployments, database migrations, frontend builds, rollback capabilities
• Scale: Manages infrastructure for 10+ production applications with zero-downtime deployments

Production multi-provider LLM system with ML models for entity deduplication at scale

Enterprise AI system for validating and deduplicating hotel property data at scale across multiple supplier datasets.

GenAI & Agentic AI Architecture:

• Multi-provider LLM abstraction layer — LLMIdentityProvider protocol with Claude (Anthropic), Gemini (Google), and OpenAI as interchangeable backends for hotel identity verification
• MultiProviderIdentityService — fallback chain + second-opinion consensus. When Claude and Gemini disagree, OpenAI arbitrates
• Agentic pipeline orchestrator — autonomous multi-step workflow: enrich data → validate against MongoDB → score → flag anomalies → notify
• Cost-aware LLM management — token budgeting, daily spend caps, call metrics, provider cost tables, retry logic

ML / Data Science:

• LightGBM gradient boosting model — trained for false-positive filtering with group-held-out cross-validation (ROC-AUC: 0.68)
• Record linkage scoring — fuzzy string matching on name/address/phone/website + geospatial signals using the recordlinkage library
• EnsembleOutlierDetector — combines LightGBM, record linkage, LLM identity, and canonical facts into a unified audit verdict
• Feature pipeline (lgbm_features.py) shared between training and serving to prevent train/serve skew

Backend & Infrastructure:

• Multi-stage data pipeline: MySQL sync → MongoDB enrichment → async scan → per-component validation scoring
• Microsoft Authentication (MSAL/OAuth2) for enterprise SSO
• Scheduler service for automated pipeline execution

Production-ready trading dashboard with AI-powered portfolio analytics

• Tech Stack: Python 3.11+, Streamlit, OpenAI GPT-4, Zerodha Kite Connect API, TradingView, MySQL
• Features: Real-time portfolio analytics, AI investment recommendations, professional trading interface
• AI Integration: GPT-4 powered analysis, portfolio optimization, risk assessment, market sentiment
• Trading: TradingView charts with 100+ indicators, real-time WebSocket streaming, quick buy/sell actions
• Security: Zerodha OAuth 2.0 integration, encrypted credentials, session persistence
• Export: Professional PDF reports, Excel exports, automated report scheduling

Comprehensive multi-tenant real estate marketplace platform

A full-stack real estate platform featuring property listings, user management with role-based access, document verification workflows, and multi-language support.

Key Features:

• Multi-role user management (Admin, Agent, Customer)
• Property listings with location hierarchy (Country/Province/City)
• Document upload and verification workflow
• Multi-language support (English/Spanish)
• Interactive map integration for property locations
• OAuth social authentication (Google)
• Automated email notifications

Technical Implementation:

Backend:

• Framework: Django 5.2 + Django REST Framework 3.16
• Language: Python 3.13+
• Database: PostgreSQL (production) / SQLite (development)
• Authentication: JWT (djangorestframework-simplejwt) + Djoser
• Social Auth: social-auth-app-django for OAuth integration
• Caching: Redis with django-redis
• File Storage: AWS S3 (production)
• WSGI Server: Gunicorn

Frontend:

• Framework: Next.js 15.4 with React 19
• Language: TypeScript
• State Management: Zustand 5.0
• Data Fetching: TanStack React Query 5.85
• Forms: React Hook Form 7.62 + Zod validation
• Internationalization: next-intl 4.3
• Maps: Mapbox GL 3.14 / MapLibre GL 5.6
• Styling: Sass
• OAuth: @react-oauth/google

Architecture:

• RESTful API with versioned endpoints (api/v1/)
• Multi-app Django structure (users, properties)
• Separate landing page, user portal, and admin dashboard
• Automated CI/CD with GitHub Actions

Comprehensive development and debugging toolkit

• Tools: JetBrains dotPeek, dnSpy, Database engines, Migration utilities
• Purpose: Code analysis, reverse engineering, database management, debugging
• Utilities: Access database engine, MySQL to SQL Server migration tools
• Databases: SQL Server, MySQL, PostgreSQL, SQLite, MongoDB support

Back to Top

Back to Top

Not just using AI tools — building AI systems. Multi-provider LLM pipelines (Claude + Gemini + OpenAI with consensus logic), LightGBM ML models trained and deployed in production, record linkage for entity deduplication at scale, and agentic orchestration pipelines that run autonomously.

From database design to user interface, I handle every aspect of application development with equal proficiency—complete top-to-bottom execution.

Deep expertise in both AWS and Azure cloud platforms—designing, deploying, and managing scalable cloud infrastructure and services.

Designing end-to-end enterprise solutions with focus on scalability, performance, and maintainability—turning complex business requirements into robust technical architectures.

Successfully led development teams of 5-15 developers, mentored junior developers, and managed end-to-end project delivery with Agile methodologies—ensuring on-time, on-budget execution.

Back to Top

Back to Top

I'm passionate about transforming complex business requirements into elegant, scalable solutions. Whether you need:

• Enterprise Architecture - Designing systems that scale
• API Development - Building robust, documented APIs
• Travel Technology - Booking systems and integrations
• Cloud Migration - Moving to Azure/AWS
• Team Leadership - Leading development initiatives
• Legacy Modernization - Upgrading existing systems

I deliver enterprise-grade solutions with clean code, comprehensive documentation, and measurable business impact.