I wire LLMs to real products, data, and users.
I'm Hanna Hinn, a backend-leaning full-stack engineer in Ramallah, Palestine. I lead engineering at Callifly, where I ship a production AI voice & chat platform serving 100+ users and 50+ maintained AI agents on a Terraform-managed AWS serverless stack. On the side I build VectorForge, a standalone RAG engine.
Six things I've built — by what they prove, not by date.
From first commit to leading engineering on a production GenAI platform.
Lead Full-Stack Engineer · Callifly · France (remote)
Engineer AI-powered voice and chat assistants end-to-end. Designed the GenAI integration layer connecting LLMs to internal knowledge bases & external platforms (e-commerce, CRM) via custom tools and APIs. RAG cut customer response times by ~50%; CI/CD & automated alerting cut manual ops by 30%+. Collaborate directly with the CTO on architecture.
Back-End Engineer · ImVitro · Paris (Biotech / IVF SaaS)
Hardened serverless reliability — fixed bugs, expanded unit-test coverage, updated infra templates. Built error-handling pipelines (SNS → Slack) that reduced operational cost. Shipped report-generation backend with S3 & new APIs / Lambdas / API Gateway endpoints to support evolving product needs.
Software Development Trainee · TAP · Nablus
Talent Acceleration Platform. Built real-time web services with Node, Express, MongoDB; learned to secure data to OWASP standards; rotated through code reviews and pairing with senior engineers; sharpened communication with certified career coaches.
B.Sc. Computer Science · Birzeit University
Four years of CS fundamentals — algorithms, OS, networks, databases, software engineering — alongside Java, Android, ASP.NET Core, Spring Boot, and a Django bootcamp with Gaza Sky Geeks. Side repos: PlayFair-LFSR cipher, doubly-linked-list studies, problem-solving in Java, ASP.NET Core API.
Front-End Developer Trainee · EXALT Technologies
First professional environment. Shipped a car-rental management web app with React + Firebase auth + real-time data sync; iterated on it for clean code and performance; learned to translate mockups into production while keeping UX integrity.
Backend engineer. Generative AI. AWS serverless.
I work the seam between backend systems and applied AI. Most days that means designing retrieval pipelines, wiring LLMs into product surfaces, and keeping the AWS bill honest. I came up through Birzeit, EXALT, the Gaza Sky Geeks Django bootcamp, and TAP — and have been shipping a steady stream of side projects ever since, most recently focused on retrieval and agent skills.
At Callifly I lead end-to-end GenAI feature work: from architecture review with the CTO down to React dashboards customers use to monitor every call. The platform serves 100+ users and maintains 50+ AI agents on a fully serverless AWS stack defined in Terraform.
I prefer boring infrastructure — clear contracts, reproducible builds, observable defaults — so the model, the retrieval, and the user get the attention. Off the clock I write skill prompts for AI coding agents, contribute to open source, and read about retrieval evaluation.
Notes on retrieval, agents, and shipping from outside the usual hubs.
Built in the open. Talked about when there's something to say.
VectorForge
A standalone, high-performance RAG engine. Built to make retrieval a first-class, replaceable component instead of a vendor lock-in.
AI Skills Library
A growing, evaluated catalogue of reusable skills, prompts, and patterns to make AI coding agents more reliable.
Boring infrastructure for interesting models
A walk-through of the architecture pattern behind Callifly's voice/chat AI: TS Lambdas, DynamoDB, RAG, and Terraform — and what we left out on purpose.
Stack Overflow & reviews
Answer questions when I can, run small code reviews for early-career engineers in Ramallah, and contribute notes back upstream when I find rough edges.
What I'm doing, reading, and typing on this week.
Now
Shipping RAG eval harness v0.5 in VectorForge
Leading agent-tool spec at Callifly
Learning Rust for hot retrieval paths
Open to senior backend / AI roles & chats
Reading list
Now Designing Data-Intensive Apps
Next The Pragmatic Engineer (Orosz)
Reread A Philosophy of Software Design
Papers Lost in the Middle · ColBERT
Uses
Editor VS Code · Cursor · Vim bindings
Shell zsh · oh-my-zsh · tmux
Cloud AWS · Terraform · GitHub Actions
Browser Arc · Firefox dev
Email is the fastest signal. Or book 15 minutes.
Got a hard backend, RAG, or voice-AI problem?
I'm open to senior backend / applied-AI roles, short-form RAG & LLM consulting, and a friendly 15-minute intro call. Email is fastest — usually a reply within a working day.