Prajwal HD
Prajwal HD

Hi, I'm Prajwal

An AI engineer based in California. I build AI agents and the developer tools around themProjectsAll projectsAgent tooling and developer tools I have built, with live demos and writeups., right now emberdemberdemberd.hdprajwal.devRun untrusted AI-agent code in isolated microVMs you host yourself., a runtime that runs untrusted AI-agent code inside isolated microVMs. It's open source and self-hosted, so the code runs on your own machine inside your own network.

Before this I spent three years at OpslyftOpslyft2021 to 2023Cloud Engineer, and before that Software Development Engineer, on cloud cost and reliability. on cloud cost and reliability systems, then a Master's on adversarial ML and Android malware detectionPeer-reviewed paperElectronics, 2026Re-Evaluating Android Malware Detection: Tabular Features, Vision Models, and Ensembles.. At Opslyft that work saved enterprise AWS customers about $100K a month.

I'm open to full-time AI engineer, backend, and full-stack roles. If that sounds like your team, email me or book a call.

ResumeView projectsEmailBook a call

Projects

All projects
Open FrameMake slides, carousels, thumbnails, and OG images with React and a coding agentopen-frame.dev
BaitBenchMeasures how often an LLM, writing a job application for you, falls for a prompt injection hidden in the job postbaitbench.hdprajwal.dev
emberdRun untrusted AI-agent code in isolated microVMs you host yourself, inside your own networkemberd.hdprajwal.dev
codexpassUse your Codex login anywhere, no separate API key to set upcodexpass.hdprajwal.dev

Research

Re-Evaluating Android Malware Detection: Tabular Features, Vision Models, and Ensembles

Hosahalli Dayananda, P.; Chen, Z. · Peer-reviewed, Electronics (MDPI), 2026

Compares tabular features, visual byte-plot images, and ensemble strategies for Android malware detection. Evaluates where modality choice changes the detection surface and where ensembles actually pay for themselves.

Read paper

Writing

All posts

When prompts replace code

· 4 min read

GuardScribe: Real-time speech-to-text with toxicity filtering

· 5 min read

Stack

Languages
  • Python
  • TypeScript
  • Go
  • JavaScript
AI & ML
  • PyTorch
  • scikit-learn
  • Hugging Face
  • Model Context Protocol
Backend & Data
  • Node.js
  • Bun
  • PostgreSQL
  • SQLite
Cloud & Infra
  • AWS
  • Google Cloud
  • Azure
  • Cloudflare Workers
  • Kubernetes
  • Firecracker
  • Linux
Frontend
  • React
  • React Native
  • Next.js
  • Tailwind CSS
  • Electron
  • Astro
Workflow & AI
  • Claude Code
  • Codex
  • Gemini
  • v0
  • Git
  • GitHub
  • GitHub Actions
  • Docker
  • Vercel
Analytics
  • PostHog

Talks

Get in touch

Email is the fastest way to reach me, or you can book a call.

Command Menu

Quick navigation and actions