Posts

If you use AI coding assistants, you know the cycle: paste a rough feature description into a chat box, get back three hundred lines of plausible code, and spend the rest of your day hunting down broken imports, missed edge cases, and silent regressions. The model isn’t necessarily bad at writing syntax. The problem is asking it to jump straight from a vague idea to implementation. Without clear boundaries, the model has to invent everything in between—state management, error handling, data flow, and all your unstated constraints. ...

Ask ten engineers how to structure source code, and you will get eleven different answers. The debate between monorepos and multi-repos has run for over a decade. But when you are coordinating AI agent workflows, strict data compliance, and multi-tenant IP boundaries, both models show real cracks. That is why we settled on a third option: the meta-repo. At Aliveness Ventures, we build software in regulated environments where systems have to move quickly under strict compliance rules (HIPAA, GDPR, SOC 2). On top of that, our clients own their source code, infrastructure, and documentation from day one. ...

If you are a founder building with AI, you have probably heard “guardrails” and “harnesses” as interchangeable buzzwords. They are not, and confusing them burns seed runway fast. Guardrails tell an AI what it must never do. A harness is the machinery that allows it to finish real work. Demos look flawless with clean prompts. Production is ruthless. When an agent triages support tickets, issues refunds, or queries databases, your 3:00 AM outages are almost never prompt errors. They are failures in the unglamorous plumbing around the model. ...

The No-BS Engineering Dashboard for Early-Stage Startups You’ve just hit 15 engineers. The product is finding market fit, the codebase is growing faster than your test coverage, and your CEO just Slack’d you: “Hey, we need better visibility. Can we get an engineering dashboard going?” Your first instinct might be to panic-buy an expensive analytics tool, integrate Jira with Tableau, and start tracking every single commit, PR, and story point. ...

Composition over inheritance in Go: building clear & flexible code Coming from object-oriented languages, Go’s absence of inheritance can feel like a limitation. Where were my class hierarchies? How would I model “is-a” relationships? But after time of building production Go systems, I’ve come to appreciate the clarity that comes from Go’s deliberate design choice: ditching class-based inheritance entirely in favor of composition through struct embedding and interfaces. Let me show you why this matters for your codebase. ...

Why 90% of clinical AI startups are regulatory timebombs; and the architecture you actually have to build. Every week, another seed-stage healthtech founder pitches an “AI copilot for clinicians.” They show a slick demo where a doctor asks the AI a complex patient question, and the app pops out a perfectly formatted, confident answer. When I look under the hood to see how they built it, 90% of them are using standard RAG (Retrieval-Augmented Generation) wrapped around an OpenAI or Anthropic API. ...

This is an in-depth series on how to build an Electronic Health Records (EHR) System MVP. In this tutorial we’ll design and implement a vertical slice (frontend, backend, infra) of an EHR system from scratch (with enough context to understand how the pieces fit together), specifically the Appointment Scheduling feature between Patients and Practitioners. Tutorial organization This series is divided into the following parts: Introduction: Overview of the series and key concepts. Planning & Discovery: Event Storming, Example Mapping, and Domain-Driven Design. UI/UX Design: Wireframing, Prototyping, and User Testing. System Design: Architecture, Database, Security, and Integration. DevSecOps: Infrastructure, CI/CD, Observability, and Security. Frontend Implementation: Component Design, State Management, and User Interaction. Backend Implementation: Data Modeling, API Design, and Clean Architecture. Launch: Launch plan, retrospective, and next steps. Note: Each part builds upon the previous one, providing a holistic view of the EHR system development process. Feel free to jump to the part that interests you the most or follow the series sequentially for a complete understanding. ...