Software Engineer — Flutter · Applied AI · Geospatial Systems

Karachi, Pakistan • Open to Full-Time & Remote Roles

I design and build production-grade software across cross-platform application engineering, applied AI, and geospatial analytics. My focus is taking high-friction, unglamorous problems—like NP-hard academic scheduling or village-level water access quantification—and building deterministic, scalable systems to solve them.

• Architecture First: Layered separation, immutable state, testable boundaries (Riverpod, Clean Architecture).
• Grounded AI: Determinism at the core, intelligence at the edges. Using RAG, agentic workflows, and structured tool-calling over raw LLM outputs.
• Offline & Multi-Target: Offline-first state persistence (Hive/Dio caching) targeting Web, Android, iOS, Windows, macOS, and Linux from a single Dart codebase.

• Languages & Frameworks: Dart, Flutter, Python, FastAPI, TypeScript
• State & Architecture: Riverpod, Freezed, GoRouter, Clean Architecture
• AI & Data Pipelines: Agentic Workflows, RAG, Gemini API, Parallel Agents, Pandas, NumPy
• GIS & Remote Sensing: Google Earth Engine, GeoPandas, QGIS, Sentinel-2 Analysis
• Backend & DevOps: Firebase, LiveKit (WebRTC), Docker, REST & GraphQL APIs, GitHub Actions

• Problem: Open source onboarding fails at issue discovery due to high filtering friction on GitHub.
• Solution: A multi-target Flutter companion app integrating GitHub REST/GraphQL APIs with Gemini guidance to surface beginner-friendly issues with codebase orientation and suggested approaches.
• Tech: Flutter Riverpod GoRouter Dio Hive Gemini API

• Problem: University scheduling is an NP-hard problem taking weeks of manual coordination.
• Solution: Formalized timetable generation into a Constraint Satisfaction Problem (CSP) solver utilizing constraint propagation and backtracking search to produce zero-conflict schedules in minutes.
• Tech: Flutter Dart CSP Solver Firebase Riverpod

• Problem: Lack of granular, village-level data on drinking water accessibility in Sindh, Pakistan.
• Solution: Research-grade geospatial pipeline combining Sentinel-2 satellite imagery, Google Earth Engine, and GeoPandas distance-decay modeling across 5,159 settlements, validated against government datasets (UN SDG 6 alignment).
• Tech: Python Google Earth Engine GeoPandas QGIS Remote Sensing

• Problem: Measuring physical healthcare access gaps in rural underserved regions.
• Solution: Spatial accessibility index modeling facility reachability, travel boundaries, and population coverage gaps to support evidence-based public health interventions.
• Tech: Python GeoPandas Spatial Analysis Healthcare GIS

• Problem: Candidate resumes fail automated screening (ATS) due to poor keyword alignment and ungrounded formatting.
• Solution: End-to-end career copilot providing ATS scoring, grounded resume rebuilds, cover letter drafting, and interview prep using retrieval-grounded AI chat.
• Tech: Flutter Google Gemini Firebase RAG Material Design

• Problem: Technical documentation becomes outdated as codebases evolve rapidly.
• Solution: Multi-provider documentation generator orchestrating parallel AI agents to convert source code into structured API references and architectural overviews.
• Tech: Flutter Dart Parallel AI Agents Multi-Provider LLM REST API

• Problem: Live telemetry analysis tools are typically proprietary and locked behind complex desktop interfaces.
• Solution: High-frequency live F1 dashboard displaying telemetry gauges, race control leaderboards, and season analytics powered by OpenF1 and Jolpica APIs.
• Tech: Flutter Dart Telemetry Real-Time Data Firebase Auth

• Problem: Web-based co-watching platforms lack low-latency sync combined with integrated real-time video/audio chat.
• Solution: Room-based platform allowing users to watch YouTube in sync, video chat, screen share, and play Connect 4 in real-time.
• Tech: Flutter Firebase LiveKit WebRTC Real-Time Sync

• Geospatial & SDG 6: Quantifying drinking water accessibility across 5,159 settlements using remote sensing and spatial statistics.
• Constraint Solvers: Formalizing institutional scheduling constraints into deterministic graph algorithms.
• System Design: Production Riverpod architecture, multi-platform state isolation, and offline-first data sync.