Most 16-year-olds in the UK are currently memorizing slide decks for high school computing exams or building toy projects on localhost.
I took a different route. Over the past three years, I engineered production systems, founded Vedaris (https://vedaris.co.uk/), and earned over 50 industry credentials, culminating on August 31, 2026, by becoming the youngest person to pass the AWS Certified Generative AI Developer - Professional.
Here is an honest breakdown of the technical stack, the architecture trade-offs, and how treating systems design as leverage opens doors traditional schooling never mentions.
The Core Milestone: Youngest AWS Generative AI Developer - Professional
The AWS Certified Generative AI Developer - Professional is not a vocabulary test on prompt engineering. It is an intensive, scenario-driven examination of production AI systems: latency constraints, token economics, zero-trust security boundaries, and enterprise governance.
The exam scenarios focus on production realities:
Architecting sub-second vector search across multi-tenant vector stores while enforcing tenant-level KMS envelope encryption.
Preventing adversarial prompt injections, model poisoning, and toxic output using Amazon Bedrock Guardrails.
Orchestrating autonomous agentic workflows using ReAct frameworks and AWS Step Functions with deterministic fallback states.
Optimizing token throughput, cache lookups, and inferencing costs across foundation models, including Anthropic Claude, Amazon Titan, and fine-tuned open weights on Amazon SageMaker.
Earning this credential at 16 as the youngest holder required moving far beyond tutorials and building directly under production constraints.
The Real Stack: Beyond Just Writing Prompts
Real-world AI systems fail if the underlying network, container orchestration, and security layers are weak. Rather than treating GenAI as an isolated API, my stack connects machine learning to enterprise infrastructure:
Cloud Architecture and Foundations: AWS Solutions Architect - Associate, AWS Well-Architected Proficient, and Intel Cloud DevOps.
Specialized Container and Cloud Networking: Certified Calico Operator - AWS Expert (Tigera) for zero-trust Kubernetes network policies, AWS Amazon EKS, and Aviatrix Multicloud Network Associate.
Hands-On Performance Sandboxes: Verified AWS Demonstrated credentials across Agentic AI, MLOps, Serverless, Application Networking, Incident Response, and Data Lakehouse architectures.
Defensive Cybersecurity and Critical Infrastructure: OPSWAT Introduction to Critical Infrastructure Protection (ICIP), Cisco Ethical Hacker and Network Defense suites, and Google Cybersecurity Professional.
Theoretical Foundations: Harvard CS50x, deep learning through MIT 6.S191, and reinforcement learning via Stanford CS234.
Technical Projects: Production Architecture
Certifications prove baseline technical literacy; production code proves capability. Through Vedaris (https://vedaris.co.uk/), the focus is building scalable, decoupled, and secure infrastructure.
Multi-Agent DevOps Automation Platform:
Engineered an autonomous multi-agent system coordinating specialised AI agents for code review, security scanning, and deployment workflows.
Implemented an inter-agent communication protocol using LangGraph and Groq API for collaborative task execution.
Deployed serverless architecture on AWS Lambda with DynamoDB state persistence and EventBridge event-driven triggers.
Enterprise RAG System with Semantic Search:
Built a retrieval-augmented generation system processing multi-format documents (PDFs, code, structured data) with intelligent chunking.
Implemented an embedding pipeline using AWS Bedrock Titan Embeddings for high-performance semantic search.
Designed a hybrid retrieval strategy combining semantic similarity with keyword matching for precision and recall optimization.
Deployed on Amazon EKS with autoscaling, monitoring via CloudWatch, FastAPI backend, and React frontend.
Systems Leverage: The Non-Standard Route
In the UK, the conventional career advice is completely linear:
GCSEs -> A-Levels -> 3-Year Bachelor's Degree -> 100k-plus pounds in debt and living costs -> Junior Graduate Scheme
That pipeline made sense when access to compute and high-level training was locked inside universities. Today, that model carries massive financial debt (Plan 5 student loans with 40-year repayment windows) and severe opportunity costs.
My roadmap is built on systems arbitrage:
Work While Learning: Completing an apprenticeship to satisfy UK Raising the Participation Age mandates while logging enterprise commercial mileage.
Alternative Entry to Level 7: Leveraging professional certifications and real-world system delivery to enter an NCSC-certified Master's (MSc in Cyber Security) via alternative entry routes, bypassing an expensive, redundant Bachelor's degree entirely.
Enterprise Direct Contracting: Operating through Vedaris to deliver high-compliance cloud and AI infrastructure directly to clients.
Advice for Young Builders
Move Past Localhost Early: Toy scripts running on your laptop do not teach you about network timeouts, IAM permission boundaries, or API throttling. Build in the cloud under production constraints.
Learn Networking and Security: An LLM wrapper is easily commoditized. An engineer who understands Calico network policies on Amazon EKS, least-privilege IAM, and threat mitigation builds things companies actually pay to protect.
Question the Conveyor Belt: Educational systems are built for predictable compliance rather than rapid iteration. Find the leverage points, learn the actual rules, and ship production systems.
Connect with me on Credly: credly.com/users/ved-prajapati
Learn more about our enterprise architecture work at Vedaris: https://vedaris.co.uk/
Top comments (0)