Posts

There is a large installed base of consumer GPUs, and there is a growing bill for LLM inference. The obvious trade, renting the former to pay the latter, has been attempted many times and mostly produces either a token with no callers or a box rental with no product. InferSpine is a project I am building to study the missing middle: the software control plane that turns untrusted, heterogeneous, residential GPUs into an OpenAI-compatible endpoint with routing, metering, and payout. This is a note on what that system has taught me so far, including the parts that don’t work yet. ...

When you start building a new service or product, setting up a payment integration is mostly on the critical path. Whether you are running an e-commerce storefront or building a marketplace that connects service providers with customer demand, choosing the right payment infrastructure matters. In our current stack, we use Chapa, and have designed our system with clean gateway interfaces so that plugging in another provider, such as Stripe, requires minimal code changes. ...

When someone asks me how to start learning Go, the first challenge is not what to learn but where to begin. There are many resources available, but without a structure, it becomes overwhelming. This is a list of resources I’ve personally used while learning Go and during day-to-day work. Note that this list is based on personal preference and is not guaranteed to be complete. ...

I wrapped up my role at ChipChip on April 16, 2026. For two years, I worked across backend development and operations. This is not a victory lap. It is a reflection on what that period actually looked like: integrating core services, debating technical decisions with teammates, fixing operational gaps, and learning how to ship without breaking trust. ...

In SaaS applications, authentication and authorization are critical. As your platform grows to serve multiple customers, each tenant wants to use their own identity provider (IDP), users need automatic provisioning from corporate directories, and access control must work across tenants. ...

When building services that need to run reliably, you face three fundamental problems: managing secrets securely, deploying workloads consistently, and enabling services to find each other. HashiCorp’s stack addresses these with Vault, Nomad, and Consul. This guide shows how to set up and integrate all three based on what I learned while building infrastructure for distributed services. ...

As software engineers, we often work with systems that abstract away the fundamental concepts of operating systems. We deploy containers to Kubernetes, scale web services, and optimize database queries without thinking deeply about the underlying resource management. Yet understanding these foundations becomes crucial when we hit performance walls or design systems that need to handle thousands of concurrent operations efficiently. ...

Model Context Protocol (MCP) enables AI assistants to interact with external tools and services. In this guide, we’ll build a simple calculator server that demonstrates MCP concepts by creating a tool that Claude (or other MCP clients) can use to perform mathematical calculations. ...

In late 2024, I contributed to Tinode Chat1, an open-source chat platform, by adding a link preview feature2. This post highlights implementation details and lessons learned from collaborating with the Tinode community. ...