Muhammad Junaid Ejaz

SRE in the DevOps team, looking after AWS infrastructure across several accounts.

• Migrated around 80 legacy Auto Scaling Groups from Launch Configurations to Launch Templates, developing the changes as code in AWS CDK and deploying through CloudFormation change sets to avoid stack drift. Decommissioned about 20 unused groups after confirming with their owners.
• Triaged two critical PHP vulnerabilities (CVSS 9.2) across 46 flagged assets and patched the production Nagios server straight after Red Hat released its fix, coordinating the change with the owning team.
• Removed public access from 19 SNS topics across two data lakehouse accounts, clearing a critical Security Hub finding.

Outsourced to WasteHero, a waste-management SaaS platform, to build direct integrations with the archiving and ERP systems used by Norwegian municipal waste companies.

• Replaced third-party integration middleware with direct integrations to ACOS WebSak (archiving) and Visma ERP (invoicing), cutting integration costs and making the flow between WasteHero and customer systems seamless.
• Took each integration from proof of concept to production as FastAPI microservices connected to the Django and Celery backend.
• Built automatic archiving of customer-service cases under Norway's Archive Act, generating PDF letters for 9 document types and uploading them through the NOARK 5 API.
• Built scheduled Visma EDI exports over FTP, made safe against duplicate or partial uploads, concurrent runs and retries, and fixed a bug that would have sent empty invoice files to the customer's ledger.
• Wrote 166 automated tests plus architecture docs and design-decision records for both integrations.

• Built and ran containerised microservices (Docker, Kubernetes) and event-driven data pipelines processing IoT device data in near real time.
• Developed Node.js and Python backend services for secure device-to-cloud communication.
• Designed and deployed Kubernetes infrastructure for IoT platforms handling real-time device communication, managing staging and production clusters with Rancher.

• Built a real-time AI voice agent using OpenAI APIs, Whisper and WebRTC, with low-latency streaming and LLM function calling into backend services and third-party APIs.
• Added semantic search over user documents with Pinecone, fed by an OCR-to-embeddings pipeline on Azure Computer Vision and AWS Lambda.
• Designed cloud-native microservices in Node.js (TypeScript) and Python (FastAPI) on AWS and Azure, with REST APIs, JWT authentication, RBAC and secure session handling.
• Built event-driven pipelines with Kafka for real-time data and asynchronous service communication.
• Set up CI/CD with Docker and GitHub Actions, and added observability (Sentry, logging, alerts) to keep production reliable.
• Led technical design discussions, mentored engineers and contributed to architecture decisions for distributed systems.

• Led the backend migration from a legacy system to Node.js (TypeScript) microservices, improving scalability and maintainability.
• Built REST APIs for authentication, subscriptions, payments and user activity, including Stripe integration with webhook handling, retries and idempotency.
• Reduced latency with Redis caching, PostgreSQL indexing and query tuning, and set up Sentry for error tracking.
• Cut AWS infrastructure costs by 40% through EC2 right-sizing, Kubernetes autoscaling and cloud optimisation.
• Managed Kubernetes (EKS) deployments, scaling policies and service orchestration.

• Wrote a UDP packet sniffer to consume the Pakistan Stock Exchange (PSX) market feed and publish it through Redis.
• Managed backend WebSocket sessions for real-time stock trading.
• Built a Lambda and CloudWatch Events cron job to keep user application status in line with the latest PSX data, and added Firebase user management with Amazon SES email.

• Designed REST APIs in JavaScript and Express.js, and cut response times with server-side Redis caching.
• Helped build an AWS deployment pipeline with Boto3, and built Flutter mobile apps that integrate with in-house IoT products.

• Built a ticket management system with SLA enforcement and priority routing (Django, Oracle) and a real-time data feed monitoring tool (PHP, Oracle).
• Developed an e-care portal for Malitel in Angular 8, working with product managers to turn requirements into features.

Realtime Voice Assistant

Production voice assistant for real-time conversation, backend API execution and retrieval-augmented answers using vector embeddings.

OCR-to-Embeddings Knowledge Engine

Pipeline that extracts structured data from uploaded documents and makes it searchable by meaning.

MCP-based n8n Automation Server

SSE-based MCP server for agentic workflow automation across CRM, scheduling and operations.

Urdu Sign Language Interpreter

Machine learning system that interprets Urdu Sign Language.