Skip to content
Ali Abbas

Full-stack, mobile & AI engineer · Melbourne, Australia

I turn messy information into products people rely on.

I'm Ali Abbas — a full-stack, mobile, and AI engineer drawn to the whole of a hard problem: the models and retrieval at the core, the data pipelines that feed them, and the product people actually use. I learned it by building and operating Shia Library, a production search platform — on the web and as iOS and Android apps — over a large, multilingual corpus. I'm building toward AI products that ship at real scale.

View the case studyGitHubResume
~315k
hadith records ingested
4,900+
automated tests, web + mobile
91%
claims grounded (30-q eval)
Web · iOS · Android
one product, live on every surface

Selected work

Systems tested at the boundary

Production web and mobile products, real model training, grounded retrieval, and evaluation work. Each case study shows the measured result and where it stops holding.

All work
Flagship · Search & retrieval platform01

Shia LibraryLive

Web · iOS · Android

A production digital library that grew from search into answers — and from one client into two. Diacritic-insensitive Arabic search, English full-text, and semantic vector search over a corpus of over 315,000 ingested records; a grounded "Ask" (in beta) that retrieves from the live corpus and answers with Claude's native citations, so every claim links to the exact source span or the system abstains; and one Supabase/Postgres backend serving both the web platform and offline-first iOS and Android apps. Built with row-level access control, an in-app CMS, and a reading experience tuned for dense scholarly text.

Next.js 16React 19TypeScriptSupabase / PostgrespgvectorClaude (native citations)
~315k
hadith records ingested
91%
claims grounded (30-q eval)
1,000+
web tests in CI
0.94
faithfulness (Claude judge)
Read case study
Machine learning & evaluation02

PolaritySource release

A 184M-parameter cache-equivalence model, and the scientific debugging trail that established both where it works and where it fails.

PythonPyTorchHugging Face TransformersDeBERTa-v3scikit-learn
184M
trainable parameters
2/44
Veritas false accepts
Case study
Mobile engineering03

Shia Library — Mobile AppLive

iOS · Android

The same library, offline-first on iOS and Android — hand-written Swift and Kotlin keep downloads alive under each OS's rules, and the offline store makes a half-visible book impossible. Live on the App Store and Google Play.

React Native (Expo)TypeScriptSwiftKotlinObjective-C++
3,900+
automated tests (Vitest)
iOS + Android
public, one codebase
Case study
AI / retrieval systems04

Citation-Grounded RAGLive

A retrieval-augmented answer engine where every claim is grounded in a specific source — hybrid retrieval, verifiable citations, an abstain path, all measured by an eval harness.

Next.js 16TypeScriptClaude (native citations)OpenAI embeddingsBM25 + RRF
0.91
recall@5 (hybrid)
1.00
faithfulness (Claude, 22-q)
Case study
AI infrastructure05

LLM GatewayLive

A provider-agnostic LLM gateway whose semantic cache is proven correct — precision, false-positive rate, and a CI gate most managed gateways don't expose.

Next.js 16TypeScriptClaude Haiku (intent judge)OpenAI embeddingsOpenTelemetry gen_ai.*
1.00
cache precision (guarded)
0%
false-positive rate
Case study
AI evaluation & methodology06

JudgelabLive

A reproducible lab that measures how reliable LLM-as-a-judge systems actually are — chance-corrected agreement with confidence intervals and a keyless, license-clean benchmark, not a single agreement score.

Pythonuv · Ruff · mypy (strict)pytestNumPy / SciPyPydantic · Typer
1,814
aligned judge–human comparisons
κ 0.767
GPT-4 vs human (ties excluded)
Case study
Agent memory & data systems07

Engram

A bitemporal memory store for agents that won't let late-arriving or contested facts overwrite what is currently true.

TypeScriptNode.js 22 · node:sqliteSQLite FTS5node:testGitHub Actions
0/6
stale answers · baseline 3/6
19/19
correct answers · baseline 11/19
Case study
Data engineering & LLM orchestration08

Usul PipelineLive

A resilient, cost-optimized ingestion + LLM-translation pipeline that turns scattered source texts into a clean, structured corpus.

Pythonasyncio / httpxClaude (Anthropic)Supabase / PostgresPrompt caching
220
books processed
35k+
passages translated
Case study

Try it live

Open one and push on it

Some of the systems above run in public. Use them in a browser or on your phone, then read how they work.

Capabilities

What I work on

Grouped by function, not a flat list of keywords.

Languages

  • TypeScript
  • JavaScript
  • SQL
  • Python
  • Java
  • Swift
  • Kotlin
  • Bash

Frontend

  • React
  • Next.js (App Router)
  • Tailwind / MUI
  • Accessibility
  • Performance

Mobile

  • React Native (Expo)
  • Offline-first (SQLite, MMKV)
  • Native modules (Swift · Kotlin · Obj-C++)
  • Background downloads & foreground services
  • Maestro E2E
  • Release engineering (App Store, Play)

Backend & data

  • Node.js
  • PostgreSQL
  • Supabase
  • pgvector
  • Full-text search
  • Row-level security
  • Caching

AI, ML & retrieval

  • PyTorch
  • Hugging Face Transformers
  • Supervised fine-tuning
  • Model evaluation & calibration
  • RAG
  • Embeddings
  • Hybrid & semantic search
  • LLM evaluation
  • Anthropic / OpenAI APIs

Infra & practices

  • Vercel
  • CI/CD (GitHub Actions)
  • Testing (Vitest, Playwright, Maestro)
  • Sentry (privacy-first observability)
  • Git
  • Security headers (nosniff, frame, referrer), secret scanning

Writing

Notes on the hard parts

All writing
24 Jun 2026Coverage is the ceiling: what actually moved my retrieval numbersI tried the famous RAG retrieval upgrades — MMR, RAG-fusion, HyDE, multi-query, a per-source diversity cap. Most did nothing; one made things worse. What actually worked, and why only measurement could tell.7 min18 Apr 2026Making Arabic search actually find thingsWhy a naïve query misses most of its matches in Arabic text, and how diacritic-insensitive search + a client-side mirror fixes it without slowing down.6 min10 Feb 2026A translation pipeline that survives crashesTranslating tens of thousands of passages with an LLM means long, flaky, expensive runs. Here's the checkpointing and quality-gating that makes it safe.5 min

Contact

Let's talk about retrieval, scale, or correctness.

Open to software, mobile, AI, and platform engineering internships and graduate roles. Full Australian work rights, no sponsorship required.

Get in touchali@alimuhammadi.com