Documenting 17+ years architecting large-scale, distributed, cloud-native platforms. Like a honey bee building intricate systemsβI create AI-enabled IoT solutions and enterprise knowledge platforms that process millions of real-time events across supply chain, logistics, and commerce domains.
"Like the honey bee, I transform complex challenges into elegant, productive solutions"
Principal-level architect with 17+ years designing, scaling, and modernizing large-scale distributed cloud-native platforms
Container Visibility GPS AI Bot: Delivering holistic 24-hour container views, utilization intelligence, geofence awareness, door-open anomaly detection, long-wait analysis, and predictive GPS device health
Enterprise Knowledge Bot: Enabling semantic AI search across enterprise documents and EDI files using unified vector store and RAG pipelines
FlowiseAI & LangChain: Building conversational AI assistants over live IoT streams and enterprise knowledge bases
17+ years architecting mission-critical platforms across supply chain, logistics, commerce, and telecom
WiseTech Global / Blume Global
Dell EMC
Manhattan Associates
Capgemini Consulting Services
IBM India
Hands-on experiments showcasing enterprise-grade solutions in secure commerce and intelligent analytics
Comprehensive secure retail management solution with enterprise-grade authentication and AI-powered cart intelligence
Intelligent assignment based on nearest store location, availability, and hours
Enhanced security with multi-factor authentication using Keycloak
Granular permission system for different user roles
Centralized product catalog across all locations
Multi-location configuration and controls
Real-time inventory tracking system
Enterprise-style authentication patterns with AI-powered cart management that intelligently assigns items based on store proximity, real-time inventory availability, and operating hoursβoptimizing fulfillment and customer experience.
Intelligent retail analytics system covering critical business evaluations through natural language queries
Published Technical Deep Dive
Complete technical guide covering Docker deployment, WooCommerce HPOS integration, Mistral AI configuration, and real-world troubleshooting. Includes architecture diagrams, code samples, and production deployment strategies.
Read Full Article on Dev.toEnterprise implementation with custom catalog database supporting multi-location inventory, store management, and advanced retail operations across geographic regions.
Daily insights, trend analysis, and category comparisons
Stockout detection and demand forecasting
AI-driven persona generation and RFM analysis
Price elasticity modeling and margin optimization
Discount performance and cannibalization detection
Ask complex questions in plain English
WooCommerce Version: No-code AI agent using Flowise, Mistral AI, and HPOS for natural language order queries. One-command Docker deployment with real-time database access. Read full article β
Multi-Store Version: Enterprise analytics covering 35+ retail evaluation scenarios including sales performance, inventory optimization, customer segmentation, dynamic pricing, and demand forecasting across multiple physical store locations.
Technical writing and open-source work on AI systems and guardrails
GitHub repository
Open-source project exploring guardrail mechanisms for AI/LLM systems.
View on GitHubSaga orchestration for microservices with Temporal
A Spring Boot microservices sample demonstrating the Saga pattern for distributed transactions, using Temporal as the durable workflow orchestrator to coordinate multi-step, cross-service operations with reliable compensation on failure.
View on GitHubCQRS & Event Sourcing microservice sample with Axon Framework
A sample microservice built on the Axon Framework showcasing CQRS and Event Sourcing patternsβseparating command and query models and using an event store as the system of record for auditable, replayable state changes.
View on GitHubReactive Spring Boot microservice with Vert.x, Postgres & Keycloak
A microservice combining Spring Boot with the Vert.x reactive toolkit for non-blocking I/O, backed by PostgreSQL for persistence and Keycloak for centralized authentication and authorization.
View on GitHubComposite service pattern across multiple microservices
A composite microservices project demonstrating how an aggregating service layer can compose and orchestrate calls to multiple downstream microservices to expose a unified API.
View on GitHubReactive data pipeline with Kafka, Camel & R2DBC
A reactive streaming data pipeline that uses Apache Kafka for event transport, Apache Camel for integration routing, and R2DBC for non-blocking, reactive database accessβenabling end-to-end backpressure-aware data flow.
View on GitHubLow-cost AI tool-calling agent
An agentic, low-cost tool-calling template: a small HuggingFace model (Qwen2.5-0.5B-Instruct) parses natural-language instructions inside a LangGraph state machine, which calls parameterized SQLAlchemy queries against Postgres for create/search/update operations. Fine-tuned with QLoRA + TRL SFTTrainer and deployed on Modal with persisted adapters for cheap, repeatable serving.
View on GitHubTool-calling agent API with human-in-the-loop
A multi-router FastAPI service exposing dynamic SQLAlchemy CRUD over Postgres plus two LangGraph tool-calling agents backed by OpenRouter: a plain agent that executes immediately, and a human-in-the-loop agent that pauses mutating operations via LangGraph's interrupt() for explicit approval, with Postgres-checkpointed state and auto-generated thread IDs so a paused confirmation survives restarts and resumes from any process. Containerized for deployment on DigitalOcean.
View on GitHubDistributed URL shortener system design
A system design implementation of a URL shortening service, covering key generation, redirection, caching, and scalability considerations for high-throughput read/write workloads.
View on GitHubBooking/reservation platform system design
A system design exploration of a booking platform, addressing inventory locking, concurrency control, availability search, and consistency trade-offs for reservation-style workloads.
View on GitHubMulti-channel notification delivery system design
A system design implementation of a notification service, covering multi-channel delivery (email, SMS, push), retry and rate-limiting strategies, templating, and reliable fan-out at scale.
View on GitHubRide-hailing platform system design
A system design exploration of a ride-hailing platform, addressing real-time geospatial matching, driver-rider dispatch, surge pricing, and location tracking at scale.
View on GitHubPolyglot persistence and multi-database architecture demo
A demonstration of polyglot persistence, showing how to combine relational, document, and key-value stores within a single architecture to match each workload to the right data model.
View on GitHubMicro-frontend architecture with Redux, Formik, Apollo, and Neon
A micro-frontend implementation combining Redux for state management, Formik for form handling, Apollo Client for GraphQL data fetching, and Neon as the serverless Postgres backend.
View on GitHubDistilBERT + Flan-T5 on small GPUs
A technical write-up on building a hybrid LLM service combining DistilBERT and Flan-T5, optimized to run on small GPUs.
Read on Medium