Malik Saad Hayat
Malik Saad Hayat (known online as saadhtiwana) is a Polyglot Software Engineer based in Islāmābād, Pakistan, specializing in full-stack development, backend architecture, applied AI engineering, AI agents and multi-agent systems, and DevOps.[1] He is an active open-source contributor with real-world production deployments.[2]
Saad currently works as a Polyglot Software Engineer at CCRIPT and is a co-founder of the open-source Go tooling organization GoVerta. He is an active contributor to the Cloud Native Computing Foundation (CNCF) and the Linux Foundation (LFX), with a strong focus on code quality.[3]
About
Saad focuses on the gap between a working demo and a system that holds up in production, which he considers the point where most engineering careers stall. His work centres on building software that survives real traffic, failure and scale rather than software that only looks finished.
As a polyglot engineer, Saad works comfortably across several languages and layers of the stack. On the backend he builds distributed systems and microservices in Go and Spring Boot, favouring event-driven architectures, REST APIs and asynchronous processing designed for high availability. On the frontend he works in TypeScript and React, with an emphasis on performance rather than appearance alone.
A significant part of his work is in applied artificial intelligence and agentic systems. Using Python and FastAPI, Saad builds AI agents and multi-agent systems that plan, use tools, coordinate with one another and recover from failure on their own, often through the Model Context Protocol (MCP) for structured tool use. This sits alongside retrieval-augmented generation (RAG) pipelines built on embeddings and vector databases, and task-specific model fine-tuning. Underpinning these systems is a disciplined approach to LLMOps: prompt engineering, evaluation pipelines, and active control of latency and cost, monitored before deployment rather than after an incident.
His infrastructure work relies on Kubernetes on AWS for auto-scaling, Terraform for reproducible deployments, and CI/CD pipelines with observability built in from the outset. Saad is also an active open-source contributor to the Cloud Native Computing Foundation (CNCF) and the Linux Foundation (LFX), with a consistent emphasis on code quality.
Top skills
- Go (Programming Language)
- FastAPI
- Amazon Web Services (AWS)
- Large Language Models (LLM)
- Retrieval-Augmented Generation (RAG)
Services
Alongside his full-time engineering work, Saad takes on a limited number of freelance engagements. The services he provides independently are summarised below.
Backend & distributed systems
Design and development of microservices, REST APIs and event-driven backends in Go and Spring Boot, built for high availability and scale.
AI agents & LLM engineering
AI agents and multi-agent systems, RAG pipelines, vector search, model fine-tuning and LLMOps using Python and FastAPI, with evaluation and cost control built in.
Full-stack web development
End-to-end web applications with TypeScript, React and Node.js, focused on performance, reliability and clean, maintainable code.
DevOps & cloud infrastructure
Kubernetes on AWS, Terraform for reproducible infrastructure, and CI/CD pipelines with observability built in from the outset.
Developer tooling & automation
Custom developer tools, automations and AI agents that remove repetitive work and improve delivery speed across a team.
Open-source & code review
Open-source contribution, architecture consultation and code review with a consistent emphasis on quality and long-term maintainability.
Available for freelance work. Start a conversation →
Experience
The following is a summary of Saad's professional experience, in reverse-chronological order.
Software Engineer
- Architected a multi-chain crypto deposit pipeline for Bitcoin, Ethereum, and USDT-TRC20 using BIP44 HD derivation, per-user address isolation, confirmation-gated settlement, idempotent transaction processing, and fail-closed exchange-rate validation.
- Hardened a live financial backend against 22 security and performance findings, addressing authorization boundaries, cryptographic request signing, data exposure, N+1 query paths, and distributed WebSocket state consistency.
- Re-engineered critical money paths from floating-point arithmetic to exact BigInt-based accounting, backed by mutation testing and a 26-case production regression matrix; also root-caused an OOM crash loop caused by excessive catalogue memory and state synchronization.
- Engineered the production delivery and recovery layer with PostgreSQL migration checkpoints, verified backups, rollback paths, PM2 zero-downtime restarts, Nginx, Let's Encrypt TLS, Azure Blob Storage, Python automation, and GitHub Actions.
Co-Founder
- Engineering the AI stack in the open, with CNCF.
Software Engineer Intern
- Architected domain-isolated microservices for shipments, inventory, routing, carriers, and notifications with versioned APIs, strict service contracts, and independent deployment/scaling.
- Built real-time data pipelines with WebSockets, Redis, asynchronous workers, and event streams for live driver telemetry, shipment state, inventory reservations, and ETA propagation.
- Engineered concurrency-safe workflows using PostgreSQL transactions, idempotency keys, distributed locks, optimistic concurrency, and atomic state transitions to prevent duplicate reservations and corrupted shipment state.
- Designed failure-resilient event processing with retries, backoff, dead-letter queues, webhook reconciliation, and eventual-consistency patterns across carrier integrations and fulfillment workflows.
Software Engineer Intern
- Engineered Go workers to efficiently process and convert video files into streamable segments.
- Optimized video delivery with frame skipping and intelligent caching to reduce bandwidth overhead.
- Containerized the streaming stack with Docker Compose and instrumented services with Prometheus.
- Built the pipeline end-to-end, from video processing to efficient segment delivery.
Education
COMSATS University Islamabad
Coursework and project work spanning data structures and algorithms, operating systems, distributed systems, and software engineering.
Projects and open-source presence
Saad is an engineer with real-world production deployments and open-source contributions. The card below is generated live from the GitHub REST API for the profile @saadhtiwana.
This is a portal to Saad's open-source work. Visit the full GitHub profile →
Writing
Saad writes about backend engineering, APIs and DevOps on Medium. Selected articles are listed below.
Why Your REST API Sucks (And How to Fix It)
The hard truths about API design that nobody talks about: pagination, versioning, error handling and the security pitfalls that quietly break production APIs.
I Hate Docker Compose (And You Should Too)
A contrarian take on why Docker Compose falls short for real-world local development, and what to reach for instead.
More articles are published on Medium. Read all articles on Medium →
Technical skills
| Domain | Technologies |
|---|---|
| Backend | Go, Spring Boot, REST APIs, event-driven architecture, async processing, microservices |
| Frontend | TypeScript, React, Material UI |
| AI engineering | Python, FastAPI, AI agents, multi-agent systems, agentic workflows, RAG, embeddings & vector databases, fine-tuning, Model Context Protocol (MCP), LLMOps |
| Infrastructure | Kubernetes, AWS, Terraform, CI/CD, observability, Docker, Prometheus |
| Open source | CNCF, LFX, code quality, developer tooling |
Contact
Saad can be reached by email. The button below opens a pre-filled Gmail compose window addressed to his inbox.
saadhtiwana@gmail.com
- GitHub: github.com/saadhtiwana
- LinkedIn: malik-saad-hayat
- LeetCode: leetcode.com/saadhtiwana
- Medium: @saadhtiwana
Frequently asked questions
Who is Malik Saad Hayat?
Malik Saad Hayat, known online as saadhtiwana, is a Polyglot Software Engineer based in Islāmābād, Pakistan. He specializes in backend architecture, full-stack development, applied AI engineering, AI agents and multi-agent systems, and DevOps, and is an active open-source contributor to the CNCF and the Linux Foundation (LFX).
What is saadhtiwana known for?
saadhtiwana is the online handle of Malik Saad Hayat. He is known for building production-grade backend systems in Go and Spring Boot, full-stack applications in TypeScript and React, and AI agents, multi-agent systems and RAG pipelines in Python and FastAPI. He uses the same handle on GitHub, LeetCode and Medium.
What does Saad Hayat do with AI agents?
Saad builds AI agents and multi-agent systems that plan, use tools, coordinate with one another and recover from failure autonomously, often using the Model Context Protocol (MCP) for structured tool use. He pairs them with retrieval-augmented generation (RAG), model fine-tuning and LLMOps practices such as evaluation pipelines and latency and cost control.
Where did Malik Saad Hayat study?
Malik Saad Hayat studied at COMSATS University Islamabad, where he completed a Bachelor of Science in Computer Science with an A grade.
Which technologies does Saad Hayat specialize in?
Go, Spring Boot, TypeScript, React, Python and FastAPI, plus AI agents, multi-agent systems, RAG, large language models (LLMs) and the Model Context Protocol. His infrastructure stack includes Kubernetes, AWS, Terraform, Docker and CI/CD with observability.
Is Malik Saad Hayat available for freelance work?
Yes. Saad takes on freelance engagements in backend and distributed systems, AI agents and LLM engineering, full-stack web development, DevOps and cloud infrastructure. He can be reached at saadhtiwana@gmail.com.
References
- ^ "About: Malik Saad Hayat." LinkedIn. Retrieved 27 June 2026.
- ^ Hayat, Malik Saad. "The gap between a demo and a production system is where most engineering careers stall." Professional summary.
- ^ "Contributions: saadhtiwana." GitHub, CNCF and LFX activity. Retrieved 27 June 2026.