Ridwan Hamid
Software engineer building data systems, AI products and mobile applications.
Scroll to explore
I turn ideas into working products.
From production data infrastructure to intelligent software and mobile applications.
Data Engineering · Applied AI · Full-Stack & Mobile
01 / FEATURED PROJECT
Tie the Camel
AI that knows when not to answer.
A source-linked Islamic reflection aid built for grounded responses, responsible boundaries and evaluated retrieval.
Explore case studyVisit live site (opens in a new tab)
- Safety checks before the model. Deterministic checks, then a classifier with a confidence threshold, run before any answer is generated.
- Hybrid retrieval and abstention. Full-text and vector search, fused and reranked. When the sources don’t support an answer, it is built to say so instead of guessing.
- 705-scenario evaluation harness. A resumable harness. In one 254-scenario run, retrieval misses fell from 3 to 0 while median latency rose from 14.4 s to 17.3 s.
02 / FEATURED PROJECT · 2024–2025
Llongterm
Giving stateless AI a memory.
Persistent memory infrastructure, agentic retrieval and evaluation for long-running AI applications.
- Minds. Persistent AI entities. Each Mind takes in what it is told, structures it and recalls it across sessions.
- Agentic memory. Agents choose when to read, write and update. In a multi-turn task, the agent decides when to query its memory, add to it or revise it.
- Evaluation. Memory-quality scores improved from 65% to 90%, on a repeatable framework for recall accuracy, retrieval relevance and context fidelity.
03 · LightStack Solutions · Case study
03 / FEATURED PROJECT
LightStack Solutions
From concept to shipped software.
Co-founded a studio that builds iOS and Android apps and reusable engineering building blocks.
Explore case studylightstacksolutions.com (opens in a new tab)
- Distant. LightStack’s AI relationship coach, live on the App Store. Onboarding is a question graph whose answers compile into the coach’s long-term memory. An LLM-backed conversation engine draws on that memory in later sessions.
- Glucose Tracker. A client’s app that brings glucose, meals, insulin and activity into one timeline, live on the App Store. Glucose charts drawn on the GPU with Skia: the day’s profile and the response after each meal. Two-way HealthKit and Health Connect sync with offline queues and retries, so the data stays continuous. AI meal analysis from a photo or a recipe link. 1,649 unit tests, 73 end-to-end flows.
- The Engineering Foundation. Reusable modules for what every app needs, promoted into a shared foundation once two apps need them: a private package, and a scaffolder that starts each new app on it. Authentication: Sign-in and sessions. Onboarding & consent: Question flows and consent, set per app in config. Subscriptions & paywalls: Paywalls, set per app in config. Notifications: Push notifications, set per app in config. Account deletion: Revokes the Apple sign-in token first and deletes the subscription last. Shared infrastructure: Build-time-validated config, typed client files, and CI that scaffolds a fresh app from the packed release on every change. Delivery of repeatable app builds went from months to weeks. 7 releases in 5 weeks.
04 · Flight log
04 / EXPERIENCE
Flight log
The journey so far.
Six milestones, from a degree to the current role. Where roles overlapped, a line branches off and runs alongside.
- : University of Warwick, BEng Computer Systems Engineering. Dissertation: food recognition and nutrition analysis, using segmentation and classification models on a custom-curated dataset.
- : Sigma Labs, Junior Data Engineer, training programme. An intensive applied engineering programme: four case studies across ETL pipelines, real-time analytics, machine learning and infrastructure as code.
- : Llongterm, Full-Stack Engineer. Persistent memory for LLM applications: agents that decide when to read, write and update their memory, and an evaluation framework that lifted memory-quality scores from 65% to 90%. Alongside Sigma Labs, then Qrious Insight.
- : Qrious Insight, Data Engineer. Production data pipelines and lakehouse workflows with Dagster, Polars, Apache Iceberg and AWS. Migration of 6+ production pipelines from Redshift to Apache Iceberg, with reusable, idempotent loading: a projected 60% cut in weekly batch costs. Design, delivery and production support of 10+ silver-layer pipelines, turning raw behavioural data into analytics-ready datasets. A reusable categorisation framework that cut rerun cycles from a full day to minutes, and lets non-technical users run it themselves.
- : LightStack Solutions, Co-Founder. A studio building iOS and Android apps on a shared engineering foundation. Alongside Qrious Insight.
- : Google Cloud, Customer Engineer, AI Native. London.
An engineer who loves building.
I enjoy taking ideas from concept through architecture, implementation and production. Whether it’s an AI system, a mobile app or data infrastructure, I care about creating things that actually work.
- Data engineering: production pipelines and lakehouses
- Applied AI: retrieval, memory and evaluation
- Full-stack & mobile: web and App Store products
Let’s build something great.
Have a project, an opportunity or a collaboration in mind? I’d love to hear about it.
ridwanhamid501@gmail.com