About Me

I'm a software engineer with 9+ years of experience building search platforms, distributed systems, and production AI applications.

My career has taken me from founding a technology consultancy and building products from scratch to architecting fintech platforms, working on conversational AI, and building large-scale search infrastructure at Walmart Global Tech. Along the way, I've worked across hands-on engineering, system architecture, technical leadership, and production operations.

I enjoy solving problems where software engineering, infrastructure, and intelligent systems come together. Here's a little about the work I've done and the problems I enjoy solving.

━━━━━━━━━━━━━━━━━━

WHAT I WORK ON

• Search & Information Retrieval — Vespa, Solr, indexing, relevance, ranking, and search evaluation.

• Distributed Systems — Kafka, microservices, event-driven architectures, multi-tenancy, scalability, and reliability.

• Applied AI — RAG pipelines, semantic retrieval, LLM orchestration, MCP servers, and agentic AI workflows.

• Cloud & Infrastructure — Kubernetes, AWS, GCP, CI/CD, autoscaling, observability, and production operations.

• Architecture & Leadership — System design, technical strategy, cross-team delivery, and engineering practices.

━━━━━━━━━━━━━━━━━━

A FEW ENGINEERING OUTCOMES

• Reduced search market onboarding from 90 days to 2 days through configuration-driven, multi-tenant architecture.

• Reduced CI/CD cycle time from more than 2 hours to approximately 10 minutes by replacing 32 sequential builds with parallel deployment.

• Worked on infrastructure supporting 80+ services across 6 markets and processing 1.6B+ messages per day.

• Built an MCP-based AI assistant for on-call troubleshooting that reduced MTTR by 60%.

• Helped migrate Kafka certificates across 100+ services in two days, with zero rollbacks or incidents.

• Reduced compute costs by 50% in an AI workload through event-driven processing and autoscaling.

━━━━━━━━━━━━━━━━━━

MY JOURNEY

I'm currently a Senior Software Engineer at Walmart Global Tech, working on international search infrastructure. My work spans multi-tenant platforms, indexing pipelines, search evaluation, production reliability, and cross-team engineering initiatives.

One area I'm particularly interested in is search evaluation. I've worked on approaches that combine algorithmic metrics, human judgments, relevance tagging, data sampling, web crawling, and LLM-as-a-Judge. I believe improving search requires more than looking at a few queries or optimizing a single metric. We need systematic ways to understand relevance, identify failure patterns, and measure the impact of changes.

Before Walmart, I worked at IARC, a fintech company focused on loan resolution, as a Senior Software Engineer, Architect, and Acting Technical Lead. I helped define architecture and engineering practices, worked on high-throughput systems processing 400,000+ campaigns and notifications daily, modernized application architecture, and built configurable loan-resolution workflows. I also worked on performance optimization, Kubernetes operations, and technical audit engagements.

At Yellow.ai and Richpanel, I worked on conversational AI platforms, multi-tenant microservices, authentication, integrations, and data migration workflows. At Tillpos, I built APIs and serverless components for point-of-sale systems, including a reusable DynamoDB helper package that became the organization's standard data access layer.

Earlier, I founded Platform Tech Solutions, a technology consultancy. I led a team of five engineers and delivered more than five client products, from product definition through production. These included a machine-learning-based autism detection platform, a social commerce product, and a logistics platform with live tracking and fleet management.

Being a founder taught me to look beyond the code. Building products also means understanding users, prioritizing under constraints, balancing speed with quality, and making sure the solution actually reaches production.

━━━━━━━━━━━━━━━━━━

BUILDING PRODUCTION AI SYSTEMS

Alongside search and distributed systems, I've built AI systems that combine retrieval, LLM orchestration, and agentic workflows.

My experience includes:

• RAG pipelines for document ingestion, transformation, and retrieval using Python, FastAPI, Celery, Java, Spring Boot, Kafka, and Redis.

• Semantic retrieval APIs backed by vector databases for working with unstructured information.

• LLM orchestration and adaptive model routing using LangChain, Chainlit, and LiteLLM.

• MCP servers and agentic workflows that connect AI agents to tools, operational context, and multi-step engineering tasks.

• Real-time index updates using Kafka and WebSockets.

I'm interested in the engineering challenges behind taking AI beyond a working demo: retrieval quality, context management, tool orchestration, evaluation, observability, latency, cost, and failure recovery.

For me, production AI engineering isn't just about prompting a model. It's about building the complete system around it and making that system reliable enough to use in the real world.

━━━━━━━━━━━━━━━━━━

WHAT I WRITE ABOUT

I use this space to explore the engineering problems, design choices, and ideas I find interesting.

You'll find me writing about:

• Search engineering — information retrieval, indexing, ranking, relevance, and evaluation.

• Distributed systems — Kafka, multi-tenancy, scalability, reliability, and production operations.

• AI engineering — RAG, MCP servers, agentic workflows, LLM orchestration, and context engineering.

• System design — architecture trade-offs, performance, observability, and fault tolerance.

• Engineering effectiveness — automation, developer workflows, production debugging, and operational improvements.

I prefer practical explanations over surface-level summaries. I want to understand why an approach works, where it breaks down, what alternatives exist, and how we can measure whether it actually delivers better results.

My interest in search also extends into AI. As applications increasingly rely on retrieval to provide context to LLMs and agents, search quality, evaluation, and context management become fundamental parts of building effective AI systems.

I'm still learning, experimenting, and refining how I think about engineering. Writing helps me organize ideas, challenge assumptions, and share practical lessons with other engineers.

If you're interested in search, distributed systems, production AI, or the engineering decisions behind reliable software, you'll probably find something useful here.

I enjoy building systems, understanding why they behave the way they do, and finding ways to make them better. That's the perspective I hope to bring to everything I write.