Hi, I'm Sarun Saji
Software Developer — Backend & AI
I build production systems — from transport operations serving 10,000+ daily passenger movements to a published Android voice keyboard.
Aalborg, Denmark Open to Backend, Mobile AI & AI Automation roles
About Me
Software Developer — Backend & AI
I build intelligent systems that combine robust backends with AI-powered mobile experiences.
I am passionate about voice AI, multilingual typing and self-hosted systems, and about delightful user experiences backed by clean, scalable architecture.
Based in Aalborg, Denmark, and open to Backend, Mobile AI and AI Automation roles.
Selected Work
A shipped product, a data platform and software for daily operations.
Voxly AI Keyboard
A published Android product: a keyboard you talk to. Hold the mic, speak, and your words are typed into whatever app you are already in — WhatsApp, Gmail, Instagram, anything. Speak one language, send another, across 24 languages, thirteen of which get a full keyboard in their own script. Built end to end and shipped on Google Play as a paid product with in-app purchases.
- 🎙️ Speak one language, send another — say it in Malayalam and send it in Danish, or Hindi in and English out. Voxly writes in the language you picked, directly into the field you are typing in, in plain English or a Gen Z texting tone.
- 🌐 Typed it? Translate it — type in your own language, even in English letters like Manglish, and one tap on the suggestion bar swaps it for English. Undo puts your words back. Long-press the globe to write in Hinglish, Tanglish, Danish or another language instead.
- ⌨️ A keyboard rebuilt from scratch — Android's keyboard framework has been deprecated for years, so the whole board was written new: thirteen full keyboards in their own script, Nordic and QWERTZ/AZERTY layouts, swipe typing, learned words, clipboard history, emoji and GIFs, themes that follow your wallpaper, and TalkBack accessibility throughout.
- 🛡️ You always know what you sent — when the output is a language you cannot read, Voxly shows you the English of it before it goes. Your voice is never stored and the text is never logged.
- 💳 Real money, handled properly — credits are metered by the backend, never by the client, and every Play purchase is verified server-side before anything is granted. Typing is free; a voice message or a translation spends one credit, and only when it succeeds.
- 🚀 Owned end to end — Android client, backend service, the self-managed infrastructure it runs on, the Play Console listing, privacy policy and release process. Sole developer.
30M+ rows · 116× faster dashboard query
Danish Power Data Pipeline
Production data platform for the Danish electricity market: over 30 million rows of prices, weather forecasts, CO₂ intensity and consumption, orchestrated nightly in Dagster on TimescaleDB. Every morning a LightGBM model predicts the next day's 24 hourly prices before the auction closes, scored publicly as the real prices land.
175 drivers · 10,000+ daily passenger movements
Staff Transport Platform
Driver & fleet management: a Django and React platform for Emirates staff transport — manages driver duty cards, daily head counts, bus odometer tracking, delay & breakdown reporting, STM route timetables, GPS/Salik data, and driver document profiles via Google Drive.
More Work
More experiments, tools and automation — with demos and implementation details.
Rejseplanen Transit Bot + Jev
A personal Telegram bot for Danish public transport, built on n8n and the Rejseplanen REST API. Jev, TypeSafe's System One model, picks the stop you meant from the real search results and says how sure it is, so "Banegården" no longer sends you 100 km to the wrong town.
Personal RAG System
Self-hosted Retrieval-Augmented Generation system on a private VPS. Upload PDFs and links, then ask questions through a Streamlit UI or the terminal. Gemini answers from your own documents, retrieved through a ChromaDB vector index.
Aalborg DK1 Energy Dashboard
End-to-end energy-data pipeline: a nightly n8n robot pulls DK1 power prices and wind forecasts from Energinet's API into PostgreSQL, then delivers them as a Telegram briefing and a live Plotly Dash dashboard.
Cabin Crew Trip Automation
Django-based platform for cabin crew trip automation and large-scale staff transport operations (Emirates/DNATA) with real-time scheduling, reporting, and fleet management.
Experience & Education
Experience
Backend & Automation Developer
Emirates General Transport · Emirates Airline Staff Transportation Project
- Designed and deployed an end-to-end Django + MySQL platform that digitised operations for 175 drivers and 10,000+ daily passenger movements.
- Built modular backend services with role-based access control, and data-ingestion pipelines for GPS, Salik, mileage and cabin-crew Excel data.
- Delivered real-time OTP and fleet analytics dashboards (Chart.js + DataTables) and automated Excel reporting for management.
- Added Google Drive OAuth document storage and a React 18 + Axios autocomplete frontend.
- Containerised and deployed the full application stack.
Education
Bachelor of Engineering, Computer Science & Engineering
Siddaganga Institute of Technology, VTU – Tumakuru, India.
Google Data Analytics Certificate
Comprehensive training in data analysis techniques, tools, and best practices — Google / Coursera.
My Skills
Mobile & AI
Backend & Languages
AI, Data & Automation
DevOps & Tools
Let's Talk
Hiring, a project, or a question about something here? Send a message and I will reply by email.