Full Stack Software Engineer & Applied AI Architect with a disciplined focus on engineering resilient web architectures, performant mobile applications, and on-device intelligent systems. A Computer Science graduate from UET Lahore (CGPA: 3.52/4.00), I bridge the gap between complex algorithmic research and production-grade software delivery.
- Full Stack Web Systems: Architecting high-concurrency, responsive applications utilizing Next.js, React, Node.js, Express, and PostgreSQL/MongoDB with an emphasis on modular domain-driven patterns, SSR optimization, and zero-trust authentication.
- Applied AI & Edge Intelligence: Implementing real-time on-device neural networks (TensorFlow Lite, MediaPipe, MoveNet Lightning) delivering sub-200ms inference alongside enterprise LLM multi-model orchestration via streaming Server-Sent Events.
- Cross-Platform Mobile Engineering: Building maintainable, reactive mobile architectures with Flutter and Dart, employing BLoC and GetX state machines with rigorous separation of concerns.
- Product Engineering Mindset: Shipped 5+ production systems (2 live commercial SaaS platforms and 3 Google Play Store releases), translating ambiguous requirements into scalable, fault-tolerant technical solutions with measurable business impact.
- Full Stack Software Engineer (Enterprise / Scale-up / High-Growth Teams)
- AI & Computer Vision Engineer (Applied Edge & Multimodal Systems)
- Distributed Systems & Core Product Engineering (Remote / On-site)
| Domain | Proficiency | Details |
|---|---|---|
| Computer Vision & Pose Estimation | Production Grade | Real-time biomechanical analysis via MoveNet Lightning & MediaPipe; 17 anatomical keypoint tracking with <200ms latency on edge hardware. |
| On-Device Edge Neural Networks | Production Grade | TensorFlow Lite model conversion, weight quantization (INT8/FP16), latency profiling, and embedded mobile inference pipelines. |
| Large Language Models & GenAI | Advanced | Multi-model routing (OpenRouter, Claude, GPT-4), Server-Sent Events (SSE) token streaming, prompt safety guardrails, and RAG architectures. |
| Agentic Automation & Workflows | Proficient | Multi-step agent execution via n8n, automated webhook routing, Supabase vector embeddings, and persistent transaction queues. |
| Applied Machine Learning | Proficient | Supervised classification, custom CNN feature extractors, dataset curation, Scikit-learn, NumPy, and Pandas analytical workflows. |
1. Weariq AI — Production AI SaaS Platform
A commercial full-stack AI platform built to deliver dynamic generative insights, automated workflow execution, and secure user data isolation.
| Dimension | Specification |
|---|---|
| Stack | Next.js, React.js, TypeScript, Node.js, Express, Supabase, Tailwind CSS |
| Scale | Live in commercial production at weariqai.com handling real-world recurring user traffic |
| Performance | Server-Side Rendering (SSR) acceleration yielding sub-second Time To First Byte (TTFB) |
| Security | End-to-end JWT token validation, encrypted cloud storage pipelines, and granular RBAC |
| Impact | Successfully monetized and launched; automated multi-tier AI API ingestion pipelines |
| Repository | View Source / Production Platform |
Architected a resilient MERN-pattern application leveraging Next.js SSR primitives for maximal indexing and speed. Designed decoupled asynchronous background jobs to communicate with third-party AI inference endpoints without blocking user request threads, persisting state reliably within Supabase.
2. GymMate — Real-Time AI Workout Assistant (FYP)
An edge AI mobile application performing real-time human pose estimation to prevent athletic injury, validate posture, and compute repetition accuracy autonomously.
| Dimension | Specification |
|---|---|
| Stack | Flutter, Dart, TensorFlow Lite, MoveNet Lightning, MediaPipe, Supabase, n8n |
| Scale | Multi-exercise real-time biomechanical processing across mobile device form factors |
| Performance | >85% pose detection accuracy running at <200ms on-device inference latency |
| Security | Zero raw camera telemetry leaves user devices; 100% on-device vision processing |
| Impact | Final Year Project at UET Lahore; eliminated external sensor reliance using edge computer vision |
| Repository | View Source Code |
Integrated MoveNet Lightning via an optimized TensorFlow Lite runtime on Android and iOS. Built custom vector-geometry algorithms to calculate joint angles across 17 anatomical keypoints, evaluating biomechanical posture deviations dynamically while managing cross-platform UI state through reactive streams.
3. Plant Identifier — Edge On-Device Computer Vision App
A high-throughput botanical classification engine operating completely offline on consumer mobile hardware, distributed publicly on the Google Play Store.
| Dimension | Specification |
|---|---|
| Stack | Flutter, Dart, TensorFlow Lite, Convolutional Neural Networks (CNN), Android SDK |
| Scale | Shipped and globally accessible on the Google Play Store with thousands of offline classes |
| Performance | Sub-150ms image classification pipeline using quantized neural model weights |
| Security | Standalone offline execution model ensuring absolute privacy of local media assets |
| Impact | Production Play Store release providing instant botanical diagnostics without internet reliance |
| Repository | View Source Code |
Engineered and deployed a mobile-optimized CNN model quantized for ARM NEON architecture. Constructed an intuitive camera viewport and background worker thread to prevent frame drops during camera buffer analysis, achieving seamless 60fps UI fluidity.
4. AI Assistant Web App — Enterprise Multi-LLM Workspace
A full-stack intelligent conversational interface designed for low-latency multi-model inference, session persistence, and dynamic model switching.
| Dimension | Specification |
|---|---|
| Stack | Next.js, React.js, TypeScript, Node.js, OpenRouter API, Supabase, PostgreSQL |
| Scale | Multi-turn conversational architecture managing complex nested message trees |
| Performance | Real-time token streaming via Server-Sent Events (SSE) with instantaneous chunk rendering |
| Security | Cryptographically signed session tokens, strict prompt-injection hygiene, and data encryption |
| Impact | Unified multi-provider frontier models into a single friction-free developer console |
| Repository | View Source Code |
Developed an extensible API proxy layer supporting hot-swapping between frontier language models via OpenRouter. Implemented an asynchronous PostgreSQL persistence schema that caches token usage statistics and handles message streaming with auto-reconnection recovery.
Jul 2026 – Present | Lahore, Pakistan (On-site)
Leading full-stack engineering initiatives, optimizing legacy web architectures, and implementing resilient enterprise features using modern JavaScript and TypeScript ecosystems.
- Architecting and maintaining high-performance web applications using Next.js, React, and Node.js with scalable component hierarchies.
- Designing clean, decoupled RESTful APIs and normalizing relational PostgreSQL schemas to optimize query plans and reduce database load.
- Enforcing enterprise engineering standards through code reviews, comprehensive type coverage, and integration testing across agile sprints.
Next.js React.js TypeScript Node.js Express.js PostgreSQL MongoDB REST APIs System Design
Dec 2024 – Feb 2025 | Remote
Engineered cross-platform mobile solutions with strict focus on modular code maintainability, reactive state synchronization, and native platform integration.
- Developed production-ready cross-platform mobile applications in Flutter/Dart with secure Firebase and REST API communication backbones.
- Structured codebases around the BLoC (Business Logic Component) pattern to guarantee unidirectional data flow and deterministic state transitions.
- Collaborated in an agile distributed engineering environment, maintaining clean Git feature branches, continuous integration, and rapid bug resolution.
Flutter Dart BLoC Pattern Firebase RESTful APIs Git Workflows Mobile Architecture
| Recognition | Details |
|---|---|
| Production Breadth (5+ Systems) | Independently designed, built, and shipped 2 commercial SaaS web platforms and 3 production apps live on Google Play Store. |
| Edge AI Real-Time Latency Milestone | Engineered on-device MoveNet Lightning pose estimation with >85% accuracy under <200ms latency across 17 anatomical keypoints. |
| Top Percentile Engineering Benchmark | Placed in the 92nd percentile for Software Engineering and 84th percentile for Algorithms in standardized assessment evaluations. |
| Academic & Theoretical Excellence | Graduated BS Computer Science from UET Lahore with a 3.52 / 4.00 CGPA, demonstrating rigorous technical fundamentals. |
-7c3aed?style=for-the-badge&logo=checkmarx&logoColor=white){kind=icon}
-6366f1?style=for-the-badge&logo=codefactor&logoColor=white){kind=icon}
current_focus:
learning:
- Distributed Systems Consensus & High-Throughput Event Brokers
- WebGPU & WebAssembly Accelerated Client-Side Model Inference
- Advanced Kubernetes Orchestration & GitOps Deployment Automation
building:
- Autonomous Multi-Agent Reasoning Workflows & Synthetic Data Engines
- High-Performance Biomechanical Health Analytics & Repetition Estimator
exploring:
- On-Device Quantized Small Language Models (SLMs) with Llama.cpp
- High-Concurrency Asynchronous Rust Microservices
open_to:
- Senior / Core Full Stack Software Engineering Opportunities
- AI / Computer Vision Systems Engineering Roles (Remote / On-site)
- Open Source Infrastructure & Developer Tooling Collaboration
