Mujeeb Arshad
Building intelligent systems, scalable software, and research-driven AI platforms.
I combine software engineering, artificial intelligence, and research to build reliable digital systems, experimental AI tools, and future-ready technology platforms.
About
An AI-driven software engineer building toward a research product ecosystem.
I work at the intersection of scalable software architecture, applied AI research, and human-centred intelligent systems, with a practical bias toward tools that can be used, inspected, improved, and deployed.
I am a software engineer with over six years of full-stack experience, an MSc in Artificial Intelligence from Aston University with Distinction and Best Dissertation Award recognition, and active PhD research in Agentic AI and Multimodal AI for healthcare diagnostics.
My background spans Ruby on Rails, React Native, React, PostgreSQL, MySQL, Neo4j, Elasticsearch, Docker, Python, deep learning, and research experimentation. I have worked on SaaS platforms, access control systems, graph database migration, ITSM features, mobile apps, enterprise integrations, AI research tools, retrosynthesis workflows, robotics projects, and experiment visualisation platforms.
Right now I am focused on agentic AI, multimodal fusion, respiratory sensing, and intelligent diagnostic systems, while continuing to care about the engineering fundamentals that make research products reliable: data quality, architecture, evaluation, observability, and clear interfaces.
Engineering depth
I build reliable product systems across Rails services, React interfaces, graph data stores, mobile apps, integrations, and deployment workflows.
AI systems thinking
I treat AI as a full system problem: data, models, agents, evaluation, interfaces, observability, and real-world constraints.
Research translation
My work connects experiments and applied software so AI research can become useful tools, demos, dashboards, and decision support systems.
Experience
A timeline across software products, architecture, mobile apps, and research engineering.
Drawn from my resume, LinkedIn profile, and Toptal profile: production SaaS engineering, enterprise integrations, graph architecture, AI research tooling, robotics, and PhD research.
PhD Researcher in Artificial Intelligence
Researching agentic AI and multimodal AI frameworks for healthcare diagnostics, with a focus on lung function estimation and explainable clinical decision support.
- Investigating multi-agent workflows that can collaboratively interpret physiological and respiratory data.
- Designing multimodal fusion approaches for video, metadata, and sensor-informed respiratory estimation.
- Building evaluation loops around FEV1, FVC, PEF, MAE, RMSE, R², uncertainty, and explainability.
Software Engineer
Working on production software engineering with Ruby, Rails, scalable backend workflows, and product-facing platform development.
- Contributing to production product systems with a focus on reliability, maintainability, and clean backend architecture.
- Applying full-stack engineering experience across service logic, data flows, product requirements, and delivery discipline.
- Bringing AI-aware engineering judgment into platform work, automation thinking, and system design conversations.
Developer
Built reporting, workflow, and client-critical product features across a React, Python, and PostgreSQL platform.
- Delivered log templates, major workflows, report surfaces, PDF-oriented outputs, and CPT graph features.
- Shipped urgent client-critical functionality that supported new client acquisition and product reliability.
- Fixed workflow bugs and helped onboard new development resources into the product codebase.
Postgraduate Research Assistant
Engineered AI-assisted retrosynthesis tooling for an RKE 2023 pump-priming project on converting biomass waste into biopolymer precursors.
- Built a system for product molecule input through SMILES and precursor import through structured CSV workflows.
- Implemented pathway generation and assessment logic that reports when no valid synthesis pathway is found.
- Customized Monte Carlo Tree Search selection strategy using UCT and deep reinforcement learning ideas.
- Developed reaction-SMILES logic for identifying reactive functional groups and contributed to an open-source MCTS library.
Senior Software Engineer
Led full-stack SaaS engineering across ecommerce, authorization, authentication, Rails modernization, graph data architecture, and performance work.
- Restructured group access, roles, teams, permissions, and authentication layers into a more scalable security model.
- Worked on the upgrade path from Ruby 2.4.2 and Rails 4.2 toward Ruby 3.2 and Rails 7.
- Migrated relationship-heavy product data from MySQL into Neo4j graph-backed architecture.
- Improved REST API performance by tracing and fixing slow Neo4j queries, and supported Material UI v4 to v5 migration.
Senior Full-stack Developer
Delivered authentication, authorization, graph visualisation, and scalable product architecture work for a fintech product environment.
- Designed graph-linked authorization and authentication architecture for clearer control over roles and teams.
- Created D3 graph and node visualisations for complex data relationships and product insight surfaces.
- Worked across backend, frontend, testing, responsive interfaces, observability, and architecture decisions.
Full-stack Developer
Improved SaaS payment reliability, Stripe-backed recurring payment workflows, SSL operations, and backend/frontend product fixes.
- Resolved bugs in recurring payment flows and improved billing workflow consistency.
- Fixed inconsistent Stripe user data and improved webhook-backed payment behaviour.
- Renewed broken SSL certificates and stabilized production application surfaces.
Senior Software Engineer
Worked on SaaS product architecture, enterprise integrations, CMDB visualisation, Rails modernization, and high-availability engineering.
- Developed ITSM and asset-management features across EZOfficeInventory, EZRentOut, and AssetSonar.
- Built CMDB features with D3.js and helped make connected asset and configuration data easier to inspect.
- Improved enterprise security and provisioning workflows used by major customers including Netflix, NASA, and Microsoft.
- Received the Dedicated Engineer award for product ownership and delivery impact.
Software Engineer
Built full-stack SaaS features, enterprise integrations, React Native mobile applications, access-control systems, and service desk workflows.
- Implemented integrations across SCCM, G Suite, Okta, Azure AD, Jira, Zendesk, SCIM, SAML, and LDAP.
- Built React Native and Rails mobile application surfaces from scratch for production SaaS workflows.
- Patched a critical SAML authentication gem issue and worked on authentication and authorization features.
- Helped grow AssetSonar from early subscriptions into a major ITSM platform, with Toptal profile impact noted as 5 to 250+ subscriptions over two years.
- Received the Initiative Taker award for proactively shipping product improvements.
Software Engineer
Worked on code-generator features, product architecture research, and payment integrations for application scaffolding and salon-management workflows.
- Implemented code-generator application features and researched maintainable code structures.
- Built Stripe integration work for salon-management workflows.
- Worked across Rails APIs, React, background jobs, CI, and rapid product delivery.
Volunteer Research Assistant
Supported early applied AI and IoT research work around home automation and deep learning experimentation.
- Explored IoT home automation concepts and applied deep learning research workflows.
- Connected early research experimentation with practical software engineering implementation.
- Built research habits that later shaped MSc and PhD work in applied AI systems.
AI & Research
Research systems for agentic AI, multimodal fusion, and respiratory intelligence.
The lab-facing side of the platform focuses on experiments, model evaluation, diagnostic signals, explainability, and research tools that can evolve into product-grade software.
Multimodal Research Stack
Signals
Video, metadata, respiratory traces
Fusion
Feature, decision, and agent-level strategies
Models
Baselines, deep learning, comparative evaluation
Estimates
FEV1, FVC, PEF, MAE, RMSE, R²
Explanation
Model behaviour, uncertainty, and diagnostic context
Agentic AI
Multi-agent workflows for model orchestration, decision support, research automation, and healthcare diagnostic reasoning.
Multimodal AI for Healthcare
Combining video, metadata, physiological signals, and sensor-derived evidence for more reliable respiratory estimation.
Lung Function Estimation
Estimating FEV1, FVC, PEF, and related markers through intelligent diagnostic systems.
Respiratory AI
AI methods for respiratory sensing, clinical signal interpretation, and explainable model output.
Machine Learning Pipelines
Reproducible data preparation, training, evaluation, and visual inspection workflows.
AI-assisted Retrosynthesis
MCTS-backed pathway exploration for biomass waste conversion, SMILES-based molecules, and biopolymer precursor research.
Robotics and Simulation
ROS, Gazebo, Webots, navigation, localisation, and kinematics experiments from applied AI and robotics training.
Research Visualisation
Dashboards and tools that make experiments, failures, and model differences easier to reason about.
Explainable AI
Interfaces and methods that expose why models produce estimates and where uncertainty enters.
Projects & Products
Personal projects, research prototypes, and platform experiments.
This catalogue now stays focused on my own research and personal builds. Cards only link out when there is a public demo or repository available.
Research Experiment Visualisation Lab
A dashboard-oriented lab for inspecting model result CSVs, comparing runs, and preparing research-ready figures for PhD experiments.
AI-Assisted Retrosynthesis and Biopolymer Pathway Tool
A research assistant tool for entering product molecules with SMILES, importing precursor molecules from CSV, generating synthesis pathways, and assessing biomass-to-biopolymer conversion routes.
MCTS Simple Open-source Contributions
Open-source contribution work around Monte Carlo Tree Search experimentation, UCT selection strategy, and research-oriented pathway search.
Personal NAS-hosted Research Platform
A self-hosted platform concept for private research tools, experiment dashboards, demos, and deployment-ready prototypes.
Fraudulent Activity Detection Experiments
Machine learning experiments on BankSim-style transaction data using Random Forest, XGBoost, and KNN for fraud detection workflows.
Home Service Robot
A robotics project using Gazebo mapping and autonomous navigation to move through an environment and complete pickup/drop-off tasks.
Where Am I ROS Localisation Workspace
A ROS package and custom Gazebo robot model for localisation and simulation-driven robotics experimentation.
Robotic Arm Grasping
A Webots-based robotics project exploring forward and inverse kinematics for robotic arm grasping behaviour.
RhythmHive Intelligent Music App
An intelligent music application concept that adapts playback and recommendations from facial-expression-driven emotion cues.
Research Lab
A dedicated lab for the Experiment Dashboard.
Visit the research lab for the focused experiment dashboard surface. The main platform keeps the broader profile, projects, writing, and career context here.
Skills
A practical stack for building platforms, research systems, and AI tools.
Grouped by how the skills are used, from product interfaces and backend services to machine learning pipelines and experiment visualisation.
Interfaces for platforms, tools, dashboards, and product workflows.
Writing & Notes
Selected technical writing and essays.
A small set of published Medium posts for now. This section can later grow into first-party notes, tutorials, and research logs.
Challenges We Face as Middle-Class Start-Up Hopefuls
A reflective essay on constraints of trying to build a startup from a middle-class background.
Collapsible Network Graph Diagram with D3
A technical walkthrough for building an interactive collapsible network graph diagram using D3.js.
Supporting Native Features to Expo Application Using EAS Build
A practical guide to enabling native capabilities in an Expo application using the EAS Build workflow.
Contact
Let's build something research-driven and useful.
For software platforms, AI research tooling, experimental dashboards, PhD collaboration, or product-minded research prototypes.
Primary contact
mujeebarshad62@gmail.comThe platform still reads this from environment configuration, so production contact details can change without editing application code.