Mugunth Kumar (MK)

CTO & Engineering Leader

I build AI-powered systems that scale. Most recently as CTO at Pink Cloud, I led the Australian expansion of an evidence-based addiction recovery platform from California through strategic partnerships with University of Melbourne, Monash, and government health sectors.

I’ve spent two decades at the intersection of early-stage technology adoption and hands-on engineering — from pioneering iOS development in 2008 to building modern semantic search and ML-integrated platforms today. My internationally-translated programming books (h-index: 35) reflect the same skill I apply as a leader: understanding where an engineer is, meeting them there, and accelerating them to productivity. I’ve done this at scale — through published books adopted by universities, and in every team I’ve led.

I’m a polyglot engineer — I believe a software architect should see the big picture and use the right language and stack for the right product, user, and scale. I’ve shipped production systems in Python, Rust, Go, TypeScript, Kotlin, and Swift, and I combine that technical breadth with the ability to structure teams, mentor engineers, and drive strategy across startups and scaleups. I build systems and the teams that sustain them.

Work Experience #

• Led Australian expansion for Pink Cloud’s evidence-based addiction recovery platform (gopinkcloud.com); also served as Director of Pink Cloud Australia (local subsidiary)
• Built strategic partnerships with University of Melbourne, Monash University, and Turning Point (Australia’s leading addiction research centre)
• Liaised with government contacts and academic institutions to establish research collaborations and government partnerships
• Positioned Pink Cloud’s AI-integrated platform for enterprise and government health sector deployments across Australia

• Provided strategic and architectural guidance across the full technology stack for evidence-based addiction recovery platform
• Built Subscriptions Management Platform - Django-based multi-tenant subscription system with academic partner analytics dashboard
  • Designed secure subscription code generation system with 5M+ unique combinations and privacy-first architecture
  • Implemented Celery-powered background processing for large dataset exports and automated notifications
  • Built comprehensive analytics dashboard for research organizations tracking participant retention and recovery journey insights
  • Deployed on AWS ECS Fargate with Aurora Serverless v2, Docker containerization, and CI/CD via GitHub Actions
• Built Meeting Data Platform - Comprehensive ETL system for aggregating addiction recovery meeting data
  • Architected FastAPI-based REST API serving meeting data from DuckDB with sub-100ms query latency
  • Designed extensible Python scraper framework using aiohttp and BeautifulSoup4 for multi-source data extraction
  • Implemented monorepo architecture with Docker Compose orchestration, automated testing, and pre-commit hooks
  • Deployed scraper as scheduled ECS Fargate task and API as auto-scaling ECS service with ALB, reducing operational costs to ~$30/month
• Supported product and engineering leadership on technical decision-making, team structuring, and scaling architecture for growth

• Strategic short-term contract to evaluate and integrate advanced AI-powered search platforms into the mobile app experience.
• Worked across GraphQL, TypeScript, and Swift to build integration proof-of-concepts and evaluate developer experience, scalability, and latency impacts.

• Building Nool (read.nool.au) - A reading companion platform with semantic book discovery — a capability absent from competitors like Goodreads and StoryGraph
• Semantic search: pgvector embeddings on highlights, books, and per-user centroids with cosine similarity ranking; chose pgvector over Qdrant for transactional integrity
• MCP server exposing semantic reading tools — the LLM acts as a “knowledgeable, well-read friend” using the reader’s embedded identity, not generic recommendations
• Technology Stack: Django 5.1+, Python 3.13, HTMX, TailwindCSS, PostgreSQL, Redis, Celery, Stripe subscriptions; native iOS and Android mobile apps
• Self-hosted infrastructure: Forgejo CI, Cloudflare Tunnel → Caddy, Watchtower auto-deploy, Docker on homelab
• KOReader and Audiobookshelf integrations for automatic highlight sync and bidirectional deep linking from e-readers and audiobook servers
• Embeddings journey: Started with highlight-vectors (pgvector + 7K real book highlights), then built Resonance (semantic music search with Qdrant + OpenAI embeddings), before shipping production semantic discovery into Nool with pgvector — a deliberate, ground-up approach to understanding vector search rather than plugging in an API
• Shipped Exhibit AI (v2 in progress) — multi-source evidence analysis pipeline (email, iMessage, WhatsApp) for family law proceedings, with privacy-first architecture using local LLM screening (Ollama) so raw client data never leaves the network before curated subsets reach cloud models for detailed analysis
• Building Panicade — biometric data analysis tool using Polar H7 chest strap RR intervals to map autonomic nervous system patterns; exploring HRV, vagal tone, and co-regulation visibility in heart rate data. Built with Python and real physiological data

• Led 24 engineers across 5 cross-functional teams (3 local, 2 offshore)
• Restructured team structures within 90 days, coaching tech leads into people leaders and increasing delivery autonomy.
• Introduced Story Mapping to improve cross-functional alignment and reduce engineering/PM communication overhead.
• Implemented a formal incident management process and trained associate engineers to run production releases within 3 months.
• Achieved significant reduction in cycle times and improved delivery velocity.

• Promoted from Mobile Lead to Engineering Leader within 2 months, managing a cross-functional team of 14 (React.js, .NET, iOS).
• Expanded hiring internationally (New Zealand, Colombia) during COVID, improving team diversity and global delivery capabilities.
• Drove alignment across distributed teams working on complex, multi-platform features.

• Instrumental in establishing the Swift Programming course at RMIT, recognised by Apple as part of their global “Everyone Can Code” initiative

• Delivered Xero Me mobile app within one year.
• Led the migration from Objective-C to Swift using Branch by Abstraction and feature toggles, enabling safe, incremental rollout.
• Implemented A/B testing and phased rollouts, decoupling engineering delivery from product release schedules and reducing risk.

• Founded and scaled a boutique product engineering firm in Singapore, running it successfully for 7 years.
• Led hiring and team growth from 1 to 10 engineers, serving as the primary engineering team for RedMart (Singapore’s leading online grocery platform, later acquired by Alibaba for integration into Lazada).
• Designed and implemented a scalable, BFF-style architecture for iOS, Android, and React.js frontends that supported RedMart’s rapid growth.
• Created a structured hiring rubric aligned with capability frameworks for consistent, bias-free hiring.

Technical Focus Areas #

AI/ML & Backend Development #

• Build production-grade, AI-integrated systems using FastAPI for high-performance APIs and scikit-learn for ML workloads.
• Hands-on experience with vector embeddings, Qdrant, pgvector, and cosine similarity — built understanding ground-up across three projects before shipping semantic search in production.
• Architect cross-platform AI systems integrating RAG pipelines, and XPC Services for macOS-native workload isolation.
• Experience with AWS Bedrock for scalable, cloud-based generative AI workflows.

Technical Leadership & Team Building #

• The same coaching methodology behind my published books (h-index: 35, university-adopted) is what I apply in every team — assess where each engineer is, build from their existing skills, and accelerate them to productivity and engagement.
• Turned 4 Go engineers into mobile specialists at PaperCut, eliminating external hiring needs entirely.
• Coached tech leads into people leaders within 90 days at Education Horizons; trained associate engineers to run production releases within 3 months.
• Scaled teams from seed-stage (1 engineer) to 10+ at RedMart, serving as the primary engineering team through to Alibaba acquisition.
• Built objective, bias-free hiring rubrics and capability frameworks at every company. Impact: Reduced attrition from 30% to under 5%; improved gender diversity from 5% to 30%.

System Architecture & Scalability #

• Designed event-driven architectures, Backend for Frontend (BFF) layers, and microservices for complex product ecosystems.
• Implemented Google Pub/Sub for horizontally scalable payment processing.
• Built Redis-based optimization layers and deployed Celery/Django Q2 for background workloads and distributed tasks.
• Delivered safe deployment strategies via feature toggle systems and progressive rollouts. Technologies: Event-driven systems, BFF, Redis, Celery, Django Q2, microservices.

Cloud Infrastructure & DevOps #

• Built AWS cloud infrastructure using CDK (TypeScript) with production-grade VPC, Fargate, RDS, and ElastiCache.
• Enabled observability with CloudWatch dashboards, structured logs, and real-time alerts.
• Automated CI/CD pipelines (e.g., with Bitrise) to support frequent releases by junior engineers. Technologies: AWS CDK, Docker, Fargate, RDS, ElastiCache, CloudWatch, Bitrise

Cross-functional Collaboration & Process #

• Drove Agile maturity via Story Mapping, DORA metrics, and structured incident management.
• Coordinated globally distributed teams across time zones and technical stacks.
• Designed capability frameworks, technical guilds, and objective hiring rubrics to scale knowledge and team capacity. Methodologies: Agile, Story Mapping, DORA metrics, ADRs, Feature Toggles

Technologies I Work With #

Primary Expertise #

• Python – Backend & AI/ML systems with FastAPI, Django, and scikit-learn
• Swift/SwiftUI & Objective-C – iOS development since 2008, including large-scale apps and platform migrations
• AWS – Infrastructure-as-code with CDK; services include Fargate, RDS, ElastiCache, and CloudWatch
• Golang – Microservices and Backend-for-Frontend (BFF) design for mobile-facing APIs
• Docker – Containerization and local dev automation
• Redis – High-throughput caching, job queues, and message brokering
• Rust – Linux desktop apps (GTK4/libadwaita) and systems tools — chosen for memory safety and performance where it matters

Additional Proficiency #

• Kotlin & Kotlin Multiplatform – Cross-platform mobile development
• React.js & Node.js – Full-stack JavaScript for web platforms
• Kubernetes – Container orchestration for scalable deployments
• C# .NET, Java, PHP – Enterprise backend systems and service layers
• Scala – Data processing and analytical workloads

Publications & Pedagogical Impact #

• iOS Programming: Pushing the Limits (Wiley) — a trilogy covering iOS 5, 6, and 7
  • Translated into Spanish, French, and Chinese
  • Adopted as a course textbook at the University of Portland
  • h-index: 35 — reflecting sustained influence as a practitioner-educator; these citations come from developers and universities building on the work
  • The same ability to assess a reader’s level and build understanding from where they are is what drives my engineering leadership — coaching, mentoring, and upskilling engineers is teaching, applied to teams

Author Page on Amazon

Philosophy #

I believe in Kaizen (continuous improvement) and occasional Kaikaku (radical change) when needed. My leadership style focuses on transparency, data-driven decisions, and fostering an environment where both technical excellence and human connections thrive.

Download PDF Version