HN user

iamfraol

1 karma
Posts0
Comments2
View on HN
No posts found.

Location: Addis Ababa, Ethiopia | Remote only Willing to relocate: No

Technologies: Python, FastAPI, Next.js, React, TypeScript, Tailwind CSS, Pydantic, MongoDB, Supabase, Gemini API, Framer Motion, Vercel, Railway Resume/CV: https://www.linkedin.com/in/fraol-ashebir-218b89239/ Email: fraolashebir84@gmail.com

I build AI-powered web products with typed, auditable backends. Open to remote contract or full-time work.

Briefr — AI research agent that returns cited reports in seconds. Citations pulled directly from grounding metadata, not generated by the model. Migrated from deprecated google-generativeai to google-genai mid-build and revalidated grounding behavior. Live: briefr-ai.vercel.app

AppealMD — AI agent that turns dental insurance denials into payer-specific appeal letters in under 30 seconds. Built as 3 explicit steps (classify → match payer rule → generate) instead of one monolithic prompt. Pydantic validation on every AI output stage — if a required field is missing, the API raises immediately rather than shipping a malformed letter. No PHI stored. Live: getappealmd.vercel.app

GetDodos — Full-stack tutoring platform with Stripe, session management, and document handling. Most technically complex of the three. Live: getdodos.com

I avoid abstractions I can't audit. I make product decisions, not just implementation decisions.

Location: Addis Ababa, Ethiopia | Remote only

Willing to relocate: No

Technologies: Python, FastAPI, Next.js, React, TypeScript, Tailwind CSS, Pydantic, MongoDB, Supabase, Gemini API, Framer Motion, Vercel, Railway

Resume/CV: www.linkedin.com/in/fraol-ashebir-218b89239/

Email: fraolashebir84@gmail.com

I build AI-powered web products with typed, auditable backends.

Open to remote contract or full-time work.

AppealMD — AI agent that turns dental insurance denials into payer-specific appeal letters in under 30 seconds. Built as 3 explicit steps (classify -> match payer rule -> generate) instead of one monolithic prompt. Pydantic validation on every AI output stage — if a required field is missing, the API raises immediately rather than shipping a malformed letter. No PHI stored. Live: getappealmd.vercel.app

Briefr — AI research agent that returns cited reports in seconds. Citations pulled from grounding metadata directly, not generated by the model. Migrated from deprecated google-generativeai to google-genai mid-build and revalidated grounding behavior. Live: briefr-ai.vercel.app

GetDodos — Full-stack tutoring platform with Stripe, session management, and document handling. Most technically complex of the three. Live: getdodos.com

I avoid abstractions I can't audit. I make product decisions, not just implementation decisions.