Product-Minded AI, SaaS & Mobile Engineer.

PROBLEM

High-net-worth investors needed reliable portfolio visibility, secure access, and faster decision support across web and mobile.

SOLUTION

Shipped OAuth2-secured investor workflows, real-time portfolio sync, and LLM-powered insight experiences using OpenAI API across iOS, Android, and web.

RESULT

Zero production rollbacks across 6+ consecutive sprints with a consistent monthly release cadence. Reduced UI bug reports by 35% through WCAG 2.1 enforcement.

PROBLEM

SMB users faced slow app startup and unstable sessions while handling daily credit and payment operations on low-bandwidth networks.

SOLUTION

Built offline-first sync with Firebase Realtime Database and optimised the React Native runtime via FlatList virtualisation, Hermes engine tuning, and strategic memoisation — profiled with Flipper.

RESULT

Delivered a 40% faster startup, 25% drop-off reduction, and consistent monthly releases for 3 years across 100K+ active users.

PROBLEM

Distributor ordering operations needed reliability and better UX to drive retailer engagement across Pakistan.

SOLUTION

Maintained and scaled the app with targeted UX redesigns, onboarding optimisations, and Sentry-based proactive error monitoring with structured incident response.

RESULT

30% engagement growth and 99% uptime across thousands of daily B2B distributor transactions nationwide.

PROBLEM

High-order-volume operations needed faster APIs and stable app performance across retailer, picker, and driver experiences.

SOLUTION

Built the picker app from scratch as sole developer. Contributed to driver and retailer app features alongside the engineering team. Collaborated on Node.js/PostgreSQL service optimisations and owned Crashlytics crash monitoring across all three apps.

RESULT

Cut API latency 50%, scaled 3× traffic with zero infrastructure changes, and reduced crash rate from 5% to 2% — all while processing 10K+ daily orders.

PROBLEM

Visa applicants rely on overly optimistic eligibility checkers that ignore real refusal rates — leading to wasted fees, failed applications, and false confidence going into the process.

SOLUTION

Built an AI-powered scoring engine that cross-references 17 official government sources across 10 immigration regions. A structured questionnaire captures nationality, education, finances, and language scores; the model returns ranked pathway scores (0–100) with document checklists, timeline estimates, and actionable improvement recommendations.

RESULT

Transparent, calibrated assessments grounded in real refusal data. Each result shows exact point attribution (e.g. "Master's degree: +12 pts") so applicants know precisely what to improve before committing to a costly application.

PROBLEM

On platforms like Upwork, articulate non-experts routinely beat more qualified freelancers — "better talkers beat better doers." Writing a targeted, personalised proposal for every job is time-consuming and inconsistent.

SOLUTION

Built a full-stack SaaS with real-time streaming AI that generates job-targeted proposals in under 90 seconds. The system trains on the user's past proposals for voice matching, analyses job posts to lead with the most relevant experience, and ships with a proposal pipeline tracker, integrated job feed, and freemium monetisation.

RESULT

A complete freelance productivity tool in production — with Pro tier ($19/month), A/B proposal testing, live streaming output, and style-matched personalisation that sounds like the user, not a template.

PROBLEM

Blockchain engineers are in critical global shortage, yet most learning resources are passive and theory-heavy — disconnected from real development environments and impossible to act on without external tooling.

SOLUTION

Contributed to the React frontend of an interactive Bitcoin/blockchain learning platform — working alongside the Vaionex engineering team to build and maintain course interfaces, mentor booking flows, and the in-browser coding environment.

RESULT

Platform grew into the leading Bitcoin education resource — integrated into the University of Exeter's official curriculum for 2+ consecutive years, showcased at BSV Global Blockchain Convention, and backed by Vaionex Corporation with a global developer community.

PROBLEM

Pakistan's retail supply chain was fragmented across opaque wholesale channels — retailers had no reliable way to verify supplier authenticity, compare prices transparently, or place bulk orders without physically visiting wholesale markets.

SOLUTION

Built the Next.js storefront for Pakistan's largest B2B wholesale marketplace, connecting verified suppliers with retailers across 8+ product categories. Delivered wholesale pricing (up to 30% below retail), flexible MOQs, and a Sentry-monitored checkout experience with 3-day delivery across major cities.

RESULT

A high-traffic B2B marketplace serving retail shop owners across Karachi, Lahore, and Islamabad — thousands of SKUs, a verified supplier network, and a performant Next.js frontend handling daily wholesale ordering volume at scale.