I build AI-powered applications, connecting language models, data and backend services to solve practical customer problems.
My experience spans production RAG systems, full-stack development, cloud deployment and secure software engineering. I’ve worked directly with customers to translate requirements into working software, delivered 80+ REST API endpoints, and supported applications running on AWS.
I hold an MSc in Cybersecurity with Distinction from the University of the West of Scotland and a BSc (Hons) in Software Engineering (2:1) from the University of Plymouth.
- LLM and RAG applications: Integrating language models, embeddings and retrieval into customer-facing software.
- Backend services: Building APIs, application logic and data integrations using Python, Flask, FastAPI and Java.
- Cloud applications: Deploying and supporting software with AWS, Azure, Docker and monitoring tools.
- Secure systems: Applying authentication, access control, input validation and evidence-based security analysis.
July 2023 – July 2026 · Remote
- Worked directly with customers to understand requirements and deliver full-stack applications.
- Built and deployed production retrieval-augmented generation (RAG) systems for customer-facing software.
- Developed Java, Python and Flask backend services, delivering 80+ REST API endpoints connected to SQL-backed workflows.
- Deployed applications using AWS EC2, S3, RDS and CloudWatch, supporting monitoring and production troubleshooting.
- Implemented JWT, OAuth 2.0, role-based access control and input validation.
- Tested backend services with pytest and improved code through debugging and peer reviews.
- Mentored three engineering interns over four months through code reviews, technical guidance and project work.
Centre for Defence Research & Development · October 2022 – July 2023
Contributed to an early flood-warning platform using Python and Azure.
- Connected Azure IoT Hub and Azure Functions to process environmental sensor readings at one-minute intervals.
- Validated and transformed sensor data into hourly aggregates and historical summaries.
- Supported rainfall and river-level alerts through data-processing workflows.
- Investigated abnormal readings and data-flow failures using application logs.
MSc Research Project · 2026
Built an application that combines software dependency information with public vulnerability intelligence to produce structured, evidence-backed risk reports.
Tech: Python · Flask · React · Vite · Qwen3.5-9B · PyTorch · Transformers · PEFT · LoRA
- Parses
package.jsonand lockfiles to extract dependencies and versions. - Processes public CVE intelligence into structured datasets containing affected products, severity metrics, vulnerability types and source references.
- Prepares instruction/input/output examples for a LoRA adaptation workflow.
- Integrates local LLM inference with dependency parsing, vulnerability processing and risk scoring.
- Presents affected packages, known CVEs, patch status and remediation recommendations through a React dashboard.
- Incorporates source evidence, explicit uncertainty and human review into the analysis workflow.
The research explores candidate vulnerability risks; it does not claim to predict or confirm future zero-day vulnerabilities.
| Area | Technologies |
|---|---|
| Languages | Python, Java, JavaScript, TypeScript, SQL |
| AI & Data | RAG, LangChain, Chroma, Ollama, Qwen, Llama, embeddings, pandas |
| Model Adaptation | PyTorch, Hugging Face Transformers, PEFT, LoRA |
| Backend | Flask, FastAPI, Node.js, Express.js, REST APIs |
| Frontend | React, Vite, HTML, CSS |
| Databases | PostgreSQL, MySQL, MongoDB |
| AWS | EC2, S3, RDS, CloudWatch |
| Azure | Functions, IoT Hub, App Services, Monitor |
| Delivery & Testing | Docker, Git, GitHub, CI/CD, Linux, pytest |
| Security | JWT, OAuth 2.0, RBAC, OWASP Top 10, vulnerability analysis |
- MSc Cybersecurity — Distinction
University of the West of Scotland · 2025–2026 - BSc (Hons) Software Engineering — 2:1
University of Plymouth · 2021–2024 - AI Engineer for Developers Associate — DataCamp
October 2026
Additional training in Prompt Engineering with the OpenAI API, Working with Hugging Face and LLMOps Concepts.
- Building reliable LLM applications and RAG pipelines
- Evaluating retrieval quality and model outputs
- Developing customer-facing AI solutions, from requirements to deployment
- Strengthening cloud deployment, monitoring and LLMOps practices
- Applying AI to cybersecurity with clear evidence and measurable results
- 📧 Email: supun2001hasanka@gmail.com
- 🌐 Portfolio: supunhasanka.tech
- 📝 Blog: supunhasanka.tech/blog
- 💼 LinkedIn: supun-hasanka-908741186
Building practical AI applications with secure engineering and customer needs at the centre.