AI Council
Put a controversial topic to a council of three models — they debate it live while a bigger, more capable model referees and delivers the verdict.
Launch demoAI Engineer · Freelance Developer
Adam Zailani.
I turn cutting-edge AI research into working products — on-premise LLM deployments, RAG pipelines, and computer vision systems that run on real hardware, for real clients.
currently open for freelance projects
01
About
A young tech enthusiast on a mission to make an impact — wherever I go.
I'm Adam, a freelance AI developer based in Bangi, Malaysia. With a knack for problem-solving and a passion for innovation, I bring a fresh perspective and a hunger for learning to every project I tackle — from coding complex algorithms to diving into the latest advancements in AI.
My work spans the full journey of an AI idea: researching the right architecture, prototyping fast, and then hardening it into something that runs reliably on real infrastructure — often self-hosted, always cost-conscious. I thrive on challenges and love turning ideas into reality.
02
Live Demos
Don't take my word for it — play with the work. Each demo below is a small, self-contained system I designed and built end to end.
LLM Chat
Chat with a language model hosted on my own hardware, streamed through a rate-limited proxy I built — the same self-hosted setup I deploy for clients.
Launch demoRAG Q&A
Ask questions over real documents and watch retrieval-augmented generation work — with the exact source chunks the model used, shown alongside every answer.
Launch demoHand Puppet
Puppet a procedural 3D hand with your webcam. Frames stream to my GPU host, MediaPipe extracts 21 landmarks per hand, and three.js retargets them in real time — with a skeleton overlay on your feed and a 3D puppet beside it.
Launch demomore experiments on the way →
03
Live Personal Projects
Not everything lives in a sandbox. These are real services I host on my own hardware and use every day — take them for a spin.
Jellyfin
jellyfin.arrflixx.ovh
My self-hosted media server. Requests come in through Seerr, Sonarr and Radarr hunt down the releases, SABnzbd+ downloads them, Bazarr fetches subtitles and Unmanic squeezes the files down — then Jellyfin serves it all to every screen in the house. Fully automated, zero subscriptions.
user
TestAcc pass
12345 shared tester account — locked down, be nice
Seerr
jellyseerr.arrflixx.ovh
The front door to the stack — a clean request portal where friends and family ask for movies and shows, and approved requests flow straight into the Jellyfin automation pipeline. Sign in with the same tester account and place a request yourself.
Open SeerrImmich
immich.arrflixx.ovh
My Google Photos replacement. Every shot from my phone auto-uploads to my own server, where machine learning makes it all searchable — faces, objects, places — without a single byte touching a third-party cloud.
user
demo@immich.app pass
demo the official public demo by the Immich team — my own photos stay private
it's a homelab, not a datacenter — downtime happens →
04
Experience
06/2025 — Present
Freelance Developer
Independent · Bangi, Malaysia
- Leveraged agentic coding frameworks to streamline the software development life cycle, using autonomous agents for debugging, documentation, and rapid prototyping.
- Evaluated and integrated emerging AI architectures — including RAG and agentic reasoning loops — to build smarter, more intuitive applications for diverse clients.
- Conducted continuous research into Transformer models and fine-tuning techniques so client projects use the most cost-effective, performant models available.
- Engineered and maintained Linux-based server environments with Docker service isolation and custom Bash automation for deploying AI-integrated applications.
09/2024 — 02/2025
Associate AI Researcher
MIMOS Berhad · Kuala Lumpur, Malaysia
- Engineered high-impact computer vision proofs of concept, transitioning cutting-edge research into functional prototypes for industrial applications.
- Developed robotic software within the ROS framework, bridging hardware-software communication with a polyglot C# and Python approach.
- Architected a proprietary image-generation pipeline by fine-tuning foundational models for specific in-house use cases and data requirements.
- Spearheaded on-premise LLM deployments using DeepSeek R1 — 100% data sovereignty with inference hosted on internal hardware.
- Managed technical partnerships with external engineering firms to ruggedize lab prototypes into field-deployable units.
09/2023 — 02/2024
IT Applications Intern
Edwards Lifesciences APAC · Kuala Lumpur, Malaysia
- Developed and maintained global EnterpriseOne (JDE) systems and mobile applications, ensuring seamless data flow across the enterprise.
- Migrated unsupported MAF applications to native Android, leveraging MVP architecture and REST APIs for modern performance standards.
- Resolved complex JDE tickets and deployed fixes and features to a worldwide production environment.
- Managed development tasks via Jira and executed quality audits through TestRail to maintain enterprise-level standards.
05
Skills
AI Specialization
the good stuffTools & Languages
daily driversPlatforms
home turf 06
Education & Interests
Education
Bachelor of Computer Science (Hons.)
Universiti Teknologi MARA, Jasin
10/2020 — 06/2023
3.78 CGPA Vice Chancellor's Award
Off the clock
things I'd happily talk about for hours
07
Contact
Have a project in mind, a role to fill, or just curious about AI?
say hi — I don't bite →