Saddam Hussain
Verified Expert in Engineering
React Developer
Lahore, Punjab, Pakistan
Toptal member since April 13, 2026
I work across full-stack development and AI systems, mainly building and scaling SaaS products and LLM-powered applications. Much of my work involves figuring out what’s actually breaking, whether it’s performance, architecture, or model behavior, and fixing it in a practical, production-ready way. Lately, I’ve been focused on LLM evaluation and benchmark design, improving how models are tested and how stronger, more reliable training signals are created.
Portfolio
Experience
- React - 6 years
- JavaScript - 6 years
- FastAPI - 5 years
- Next.js - 5 years
- Postman - 5 years
- Node.js - 5 years
- Django - 5 years
- AI Agents - 3 years
Preferred Environment
Visual Studio Code (VS Code), Linux, GitHub, Amazon Web Services (AWS), Slack, Claude Code, Vercel, Docker, PostgreSQL, Jira
The most amazing...
...thing I've built is a VA hiring system that automated the full recruitment flow using AI agents, reducing manual screening and coordination work significantly.
Work Experience
[Expert Crowd] Tech Support Specialist for LLM Enhancements
Vetto Inc
- Created an engineering handbook for building elite AI benchmarks, covering task philosophy, prompt engineering, trajectory analysis, 5-pillar file architecture, and verifier design to accelerate onboarding for new contributors.
- Provided technical support and guidance to 300+ contributors working on frontier AI model benchmark evaluation tasks, helping them resolve platform issues and improve task quality across multiple Slack channels.
- Identified and escalated critical platform bugs and infrastructure issues to the project manager, including evaluation report budget over consumption, smart model token loop failures.
- Proposed platform improvements adopted for client review, including a toggle to disable automatic evaluation report runs to reduce unnecessary budget consumption and a FIFO-based queue system to prioritize long-pending tasks in the reviewer workflow.
- Triaged and resolved contributor support requests across internal, external, and support channels, reducing repeated questions and noise by proactively publishing clarifying posts on recurring platform issues.
- Flagged and escalated contributor compensation concerns related to a platform-side evaluation bug that caused unintended budget per task, advocating for fair treatment of affected contributors.
Full-stack AI Developer
MEDICAL IMAGING AI PLATFORM
- Developed a mobile app and web app for a healthcare clinic based in Ireland to diagnose early symptoms of different diseases using fine-tuned versions of MedGemma 4B, and provided patient consultation through the mobile app.
- Fine-tuned MedGemma 4B for X-ray fracture detection (87% accuracy), ophthalmology diagnosis (84% accuracy), and dermatology classification (23 conditions), implemented LoRA/QLoRA with custom class weighting on resource-constrained GPUs.
- Created 30K+ medical images dataset (SCIN, AIROGS, EyePACS, FracAtlas), built an automated report pipeline processing 15K+ images with GPT-4o, handling class imbalance and validation.
- Deployed distributed inference across Kaggle/Colab with FastAPI, integrating a Django + React Native app with multi-model orchestration.
- Improved challenging class accuracy 10-15% through iterative dataset expansion (1K to 3.5K per class), learning rate tuning, and a clinical evaluation framework (sensitivity, specificity, PPV, NPV).
- Deployed distributed inference across Kaggle/Colab A100 GPUs via FastAPI for scalable model serving.
- Shipped a full Django and Next.js web app, now actively used by the clinic for patient imaging analysis internally.
Senior Full-stack AI Developer
Podding
- Built an AI-powered podcast guest booking platform solo, reducing end-to-end workflow time by 85-88%.
- Integrated a Rephonic API with filtered search to fetch thousands of podcasts by domain with episode and guest data.
- Stored podcast and guest data in a PostgreSQL and FAISS vector database for accurate semantic matching at scale.
- Built a client data ingestion module accepting any file format, auto-structuring it, and matching against podcast profiles.
- Built an automated email outreach module contacting podcast producers, tracking responses, and managing follow-ups automatically.
- Achieved a 95% podcast acceptance rate through intelligent semantic matching between guests and podcast profiles.
- Deployed LangGraph multi-agent workflows on AWS Lambda with Celery background processing, real-time analytics, and a front end on Vercel.
Founder & Full-stack AI Developer
Sitescripter AI
- Founded and shipped a commercial AI-powered Chrome extension for web productivity as a solo founder, end-to-end.
- Built a Chrome extension in JavaScript for real-time content capture paired with a Next.js dashboard and a Django back end.
- Implemented a FAISS-backed RAG pipeline enabling context-aware Q&A and personalized learning per user browsing history.
- Integrated LLMs for article summarization, intelligent Q&A, and personalized content recommendations for users.
- Deployed on AWS EC2 with CI/CD pipelines, Lemonsqueezy payments, and an admin analytics dashboard.
Full-stack AI Developer
Kayana
- Built a complete VA hiring marketplace SaaS platform from scratch working directly with the founder.
- Designed a LangGraph multi-agent system automating client onboarding, VA assessments, and meeting scheduling end-to-end.
- Implemented semantic skill matching using vector embeddings, cutting assistant-client pairing time by 85%.
- Reduced overall hiring time by 90% by eliminating manual matching through automated AI agent workflows.
- Integrated Celery task queues and background job processing with real-time notifications.
- Delivered a full application with the front end deployed on Vercel and the back end deployed on AWS.
Full-stack Developer
Revinate
- Contributed to an enterprise hotel management platform serving 5,000+ properties and processing 10,000+ daily transactions.
- Worked on React front ends and back-end services in Node.js, Python, and Java for booking management and analytics.
- Worked on real-time data pipelines using Kafka and Protobuf for booking events and customer data across microservices.
- Worked across a distributed microservices architecture, ensuring reliable inventory sync across all platform services.
Software Engineer
Kallidus
- Built Ruby on Rails REST APIs with background job processing and optimized ActiveRecord queries.
- Reduced CSV report generation from 60 minutes to three minutes by eliminating N+1 queries and implementing PostgreSQL stored procedures.
- Restructured database schema with optimized joins and strategic indexing, handling thousands of records efficiently.
- Designed a complex relational database structure with reusable stored procedures, eliminating application-level loops entirely.
- Built a custom npm state management library used across 50+ components with no Redux or MobX dependency.
- Worked on Vue to React front-end migration across the entire application.
Experience
Podding
https://podding.vercel.app/I implemented semantic matching using embeddings, allowing podcasts and clients to be scored based on context rather than keywords, improving accuracy and scalability. I also built an automated outreach system that manages communication, tracks responses, and handles follow-ups without manual work. This reduced overall workflow time by approximately 85–88% and increased acceptance rates to 95%.
Hire Kayana VA
https://hirekayana.com/I built the platform from scratch, working directly with the founder and focusing first on structuring the data and decision flow. I designed a multi-agent system in which onboarding, candidate evaluation, and matching were handled separately, rather than combining them into a single pipeline. Client requirements and candidate data were converted into structured embeddings and stored in a vector database, enabling semantic matching based on context rather than keyword filtering. This made the matching process more consistent and easier to scale.
Once matches were identified, the system automatically handled shortlisting, interview scheduling, and notifications, eliminating manual coordination. This reduced hiring time by around 90% and improved matching speed by 85% while keeping the system stable under real usage.
Mya Team Platform
https://joinmya.com/pricing-mya-team/I focused on structuring how booking availability and team roles interact at the system level. The approach was to centralize scheduling logic and enforce consistency in the back end, rather than relying on manual fixes. On the back end, I worked on APIs and data models for appointments, staff roles, and availability updates, ensuring the system remains consistent even when multiple actions occur simultaneously. This included handling edge cases around booking conflicts and real-time updates.
On the front end, I worked on building clear views for managing schedules and client data, keeping the interface simple while the system handled the complexity underneath.
Education
Bachelor's Degree in Computer Science
Comsats University Islamabad, Lahore Campus - Lahore
Skills
Libraries/APIs
React, Node.js, Zoho API, Stripe, Stripe API, Stripe Connect, NetSuite API, Playwright, API Development, NumPy, Pandas, Scikit-learn, REST APIs, Claude API, Google Drive API, Google Apps, Stripe Connect API, PyTorch, Protobuf, Socket.IO
Tools
Postman, Slack, Make (formerly Integromat), Zapier, n8n, Google Docs, Amazon SageMaker, Bower Front-end Dependency Manager, Front-end Build Tools, Git, Firebase Security Rules, Claude, Google Sheets, Mailchimp, Auth0, Claude Code, Claude Agent SDK, Adobe After Effects, Kafka Streams, Celery, Shadcn UI, GitHub, Terminal, Figma, Prisma, Jira
Languages
JavaScript, SQL, C#, TypeScript, CSS, Modern JavaScript, HTML, PHP, HTML5, Python, Java, Go, Rust
Frameworks
Next.js, Django, LangGraph, React Native, OAuth 2, .NET, .NET 9, .NET Core, ASP.NET, Ruby on Rails 5, Tailwind CSS
Paradigms
Microservices Architecture, UI Design, Visual Design, Role-based Access Control (RBAC), Real-time Systems, Automated Testing, Mobile Development, HL7 FHIR Standard, Fast Healthcare Interoperability Resources (FHIR), HIPAA Compliance, UX Design, Testing, Refactoring, Model Context Protocol (MCP)
Platforms
Zoho CRM, Android, Calendly, iOS, Docker, Firebase, Azure, Web, Hostinger, Amazon EC2, Vercel, AWS Lambda, Visual Studio Code (VS Code), Linux, Jupyter Notebook, Amazon Web Services (AWS)
Storage
PostgreSQL, Amazon S3 (AWS S3), MySQL, Redis, Cloud Firestore, PostGIS, Databases, Relational Databases
Industry Expertise
Web Design
Other
LangChain, Large Language Models (LLMs), AI Agents, Web Extensions, RAG Architecture, RESTFul APIs, Accounting Tools, QuickBooks Online, Inventory Management, eCommerce, Branding, Square, Squarespace, Business Process Automation, Customer Relationship Management (CRM), API Integration, Beta Testing, Documentation, Screen Recording, Architecture, Booking, MVP Design, Marketplaces, Scheduling, Scheduling Tools, Two-factor Authentication (2FA), User Roles, Front-end, Security, Artificial Intelligence (AI), Full-stack, Full-stack Development, AWS Bedrock AgentCore, Retrieval-augmented Generation (RAG), Data Science, Data Analysis, Finance, Open Source, Quantitative Analysis, NetSuite, Gravity Forms, AI Agent Orchestration, Pgvector, Angular 10, BullMQ, Web Scraping, Mobile App Development, Agentic AI, Process Automation, WhatsApp, Booking Systems, Scalability, Asynchronous REST APIs, Webhooks, Dashboards, Data Visualization, PDF, APIs, Blockchain & Cryptocurrency, Web Development, WebSockets, Chrome Extensions, Data Cleaning, Data Wrangling, Jupiter, Custom Keyboards, Dashboard Development, AI Automation, System Architecture, Minimum Viable Product (MVP), Technical Architecture, Senior Users, Enterprise SaaS, Supabase, Website Optimization, User Authentication, Software Architecture, Security Architecture, User Experience (UX), AI Engineering, Collaboration, Technical Leadership, Engineering, Cloudflare, HTTP, Anthropic, Deployment, Prompt Engineering, User Interface (UI), Voice, Back-end, Performance Optimization, Network Protocols, Network Engineering, Migration, Code Review, Operating Systems, Distributed Systems, Software Development, Team Management, Integration, Embeddings from Language Models (ELMo), Vector Databases, Data, AI Architecture, Natural Language Processing (NLP), AI Programming, Software Engineering, Troubleshooting, Video Editing, Augmented Reality (AR), FastAPI, Machine Learning, Pinecone, Payment APIs, CI/CD Pipelines, Background Jobs, Rephonic Api Integration, Email Automation, Open-source LLMs, GPU Computing, Fine-tuning, Google Colaboratory (Colab), AI Integration, Vectors, SaaS
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring