software · devops · ai/ml · sri lanka

jathurshan.

I build and ship reliable systems — full-stack, infra, ml, and the occasional smart contract.

open to software engineering roles · ex-IronOne·--:--:-- clt

I'm a computer engineering undergraduate at the University of Ruhuna. I recently completed a six-month software engineering internship at IronOne Technologies, where I helped build the Primary Dealer System used by Sri Lanka's fintech sector — Quarkus on the backend, React + TypeScript on the front. I'm now looking for a full-time software engineering role.

My work spans more than one lane. On the platform side: Docker, Kubernetes, AWS, Jenkins pipelines, Nginx tuning — the unglamorous wiring that keeps services alive. On the ml side: a published paper on employee-turnover prediction (ensemble KNN + Random Forest), the Stanford ML specialization, and a TensorFlow habit. And on the chain side: Hyperledger Fabric and a bit of Solidity, mostly out of curiosity for what permissioned ledgers can actually solve.

Add competitive programming on top — 19th nationally and 245th globally at IEEE Xtreme '24, 5th at HaXtreme 3.0. I care about systems that hold up under real load: secure auth, sensible APIs, observability you can actually read. Less about novelty, more about the thing not breaking at 2am.

~/whoami
nameJathurshan Thadchanamoorthyrolesoftware engineer · ex-IronOnefocusfull-stack · devops · ai/ml · web3studiescomputer eng. · UoR · 3.91/4.0backendjava · quarkus · spring · node · goinfradocker · kubernetes · aws · jenkins · nginxmlpython · scikit · tensorflow · pandaschainhyperledger · solidity · ethereumbased incolombo, lkstatusopen to swe roles · 2026emailjathurshant.eng@gmail.com
3.91
GPA / 4.0 · UoR
8,784
Teams beaten at IEEE Xtreme
23+
Certifications shipped
6 mo
Shipped fintech @ IronOne

Internships, study, and the kind of competitions that eat weekends.

IronOne Technologies
Software Engineer Intern · completed
may 2025 — nov 2025

Six-month internship in the fintech domain on the Primary Dealer System (capital-markets trading, ATrad). Built backend services in Java/Quarkus using both reactive and imperative programming, with Flyway database migrations and Keycloak authentication; shipped responsive React + TypeScript front-ends; collaborated with cross-functional teams in Agile sprints and contributed to documentation and code reviews — plus an indecent amount of debugging Keycloak.

QuarkusJavaFlywayKeycloakReactTypeScriptPostgreSQL
University of Ruhuna
Computer Engineering · undergraduate
2022 — present

BSc Computer Engineering, GPA 3.91 / 4.0. Concentrating on distributed systems, machine learning, and applied cryptography.

C++JavaPythonSystemsNetworksAlgorithms
IEEE Xtreme 24h
Competitive Programmer · 19th national / 245th global
oct 2024

Solo'd against 8,784 teams over 24 hours of algorithmic problems. Sleep, optional. Caffeine, mandatory.

C++PythonAlgorithmsProblem Solving
HaXtreme 3.0
Hackathon · 5th place
2024

5th of 93+ teams. Built and shipped a full prototype in under 36 hours.

Full StackReactNodeSpeed

A handful of things I built and shipped — or am still shipping.

2026

ChocoJava — Cloud-Native Retail Platform

Full-stack, cloud-native e-commerce platform built to stay fast, available, and secure under concurrent load. Six independently deployable Spring Boot microservices on a database-per-service model (Flyway migrations, Hibernate Envers audit trails) with a hybrid comms layer — synchronous OpenFeign behind Resilience4j circuit breakers, asynchronous Kafka event fan-out. Centralized Keycloak OAuth2/OIDC with per-service JWT validation and zero static secrets (AWS Secrets Manager + External Secrets Operator + IRSA). Deployed on multi-AZ AWS EKS with Horizontal Pod Autoscaling and NGINX Ingress/ALB, shipped via a Jenkins → ECR → Kustomize pipeline with immutable git-sha images. React 19 + TypeScript + Tailwind front with a custom design system, 8 feature pages, Radix UI, and Keycloak SSO. 4-member team · EC7205 Cloud Computing, University of Ruhuna.

Spring BootMicroservicesAWS EKSKubernetesKafkaKeycloakJenkinsReact 19PostgreSQLKustomize
eks · choco-java · multi-az
spring cloud gateway
user
product
inventory
notification
eureka
kafka
postgres ×4 · database-per-service
2026 — present

Combined Blockchain · ZKP · GNN–LLM Defense for SDVNs

Research formulating false-accusation attacks in Software-Defined Vehicular Networks as a multi-variant, cross-layer trust-manipulation problem, extending the TFMD-SDVN baseline (Nayak et al., 2022) by exposing gaps in its EMA rating, recommendation aggregation, similarity, trust-fusion, and event-validation primitives. Four attack variants — opportunistic fabrication, Sybil consensus flooding, high-noise timing, and tampered-log evidence spoofing — are formalized and simulated, then countered by a dual-mode framework: decentralized reputation on Hyperledger Fabric with weighted-endorser consensus across three smart contracts (registration auth · reputation · controller-trust audit with failover), zkSNARK accusation verification (legitimacy without identity disclosure), a temporal-attention GNN for coordinated-behaviour detection, a Transformer LLM layer for explainable incident reports over controller logs, and post-quantum primitives. Evaluated in NS-3 across 0–100% malicious-node ratios via PDR · PIR · MCC. Supervisor: Dr. P.A.D.S.N. Wijesekara, Dept. of Electrical & Information Engineering, University of Ruhuna.

Hyperledger FabriczkSNARKGNNLLM / TransformersNS-3Post-Quantum CryptoSmart ContractsPython
sdvn · combined-defense
sdvn control plane
SC1 · auth
SC2 · reputation
SC3 · audit
zkSNARK proof
GNN → LLM
PQ crypto
▦ ▦ ▦ ▦ hyperledger fabric
2026

PillPin — Real-Time Medicine-Availability Platform

Team-lead build for {PROXY}Maze '26: a real-time platform that aggregates public stock data from 3 pharmacy chains across 3 countries (Boots UK · Apollo/1mg India · Osusala Sri Lanka) into one normalized availability API and consumer map. A pluggable transport seam lets a production rotating residential/ISP proxy mesh drop in by implementing a single class and flipping one env var. Python/FastAPI backend over async SQLAlchemy 2 with an APScheduler scrape loop and a two-table freshness store; a Google Gemini AI layer normalizes messy multilingual drug names to one canonical ID and answers natural-language queries in English, Sinhala, and Tamil. Ships a metered B2B SaaS layer — JWT, RBAC, API keys, usage metering, real Stripe (test-mode) Checkout + webhooks — with a React + TypeScript consumer map and a dark B2B terminal.

PythonFastAPISQLAlchemyGoogle GeminiReactTypeScriptStripePostgreSQL
pillpin · find-a-med
rotating proxy mesh
Boots · UK
Apollo · IN
Osusala · LK
Gemini normalizer
FastAPI · /availability
postgres · current + history
2024 — 2025

DN Dental Clinic Platform

Led a 4-member team across 8 Agile sprints to build the Spring Boot (Spring MVC) backend for a dental practice — appointment booking, patient records, and payments. JWT-secured REST APIs with role-based access control powering 2 React + TypeScript frontends (patient and admin), containerized with Docker Compose and a Jenkins CI pipeline, with Prometheus + Grafana monitoring.

Spring BootSpring MVCReactTypeScriptJWTDockerJenkinsGrafana
dn-dental.lk/admin
09:00
09:30
10:15
11:00
13:45
2024

Employee Turnover Prediction

Trained and benchmarked 5 supervised models (KNN, Random Forest, Logistic Regression, Decision Tree, Gradient Boosting) on the Kaggle HR-Analytics dataset to predict turnover. Heavy data preprocessing, feature engineering, and hyperparameter tuning to surface the key retention indicators, compared on accuracy, precision, recall, F1, and ROC-AUC — methodology and findings published as a paper on ResearchGate.

PythonScikit-LearnRandom ForestGradient BoostingKNNPandasNumPy
notebook · turnover.ipynb
accuracy · 0.91epochs · 50

What's currently on the bench.

Backend

  • Java3y
  • Spring Boot2y
  • Spring Cloud1y
  • Quarkus1y
  • FastAPI1y
  • Python3y
  • Node.js2y
  • Kafka1y
  • Flyway1y

Frontend

  • React3y
  • Next.js2y
  • TypeScript2y
  • Tailwind2y

Data

  • PostgreSQL2y
  • MySQL3y
  • MongoDB2y
  • Redis1y
  • Oracle1y
  • Neo4j1y

Infra

  • Docker2y
  • Kubernetes1y
  • AWS2y
  • Jenkins2y
  • Terraform1y
  • Prometheus / Grafana1y
  • Keycloak1y
  • Nginx1y

ML / Data

  • Scikit-Learn2y
  • TensorFlow1y
  • Pandas / NumPy2y

Web3

  • Hyperledger Fabric6mo
  • Solidity6mo
  • Ethereum6mo

Where the weekends went.

Hackathon5th Place
HaXtreme 3.0 Hackathon
HaXtreme · 2024

Secured 5th place in the prestigious HaXtreme 3.0 Hackathon, competing against 93+ talented teams. Demonstrated exceptional problem-solving skills, innovative thinking, and technical expertise in a high-pressure competitive environment.

93+ teams
Competitive Programming19th Nationally
IEEE Xtreme 24-Hour Coding Competition
IEEE · 2024

Achieved 19th place nationally in the IEEE Xtreme 24-Hour Programming Competition, showcasing exceptional algorithmic thinking and endurance. Ranked 245th globally out of 8,784 competing teams from around the world, demonstrating world-class competitive programming skills.

8,784 teams globally245th Globally

Proof I kept reading the docs.

22 credentials · verified
HackerRankDec 2025
Go (Basic)
Go (Golang)
view credential
HackerRankJul 2025
Java (Basic)
Core JavaData Structures
view credential
DeepLearning.AIApr 2025
Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization
TensorFlowMachine Learning Algorithms
view credential
DeepLearning.AIMar 2025
Unsupervised Learning, Recommenders, Reinforcement Learning
Anomaly DetectionUnsupervised Learning
view credential
UdemyMar 2025
The Ultimate React Course 2024: React, Next.js, Redux & More
React.jsredux
view credential
DeepLearning.AIMar 2025
Neural Networks and Deep Learning
Artificial Neural Network
view credential
Stanford UniversityMar 2025
Machine Learning
Artificial Neural NetworkLinear Regression
view credential
LinkedInMar 2025
Learning NGINX
Nginx
view credential
Amazon Web Services (AWS)Mar 2025
AWS Cloud Technical Essentials
AWSCloud Computing
view credential

Everything else — courseware, ML experiments, side things.

yearprojectmade withlinks
2025Real-Time Network Monitoring ToolJavaJavaFXSockets
2024E-Commerce REST APISpring BootPostgreSQLJWTREST
2024Point of Sale SystemC#WPFSQLite
2024Retail Billing Automation SystemC++DSAOOP
2024Blood Bank Management SystemSQLStored ProceduresIndexing

Wall of love7 recommendations from the team I've shipped with at IronOne.

7linkedin recs
100%verified · ironone
2direct managers
6 moworking alongside
"mentor

"Delivered on time — even in the hardest situations."

Consistently impressed by his technical depth and quick learning curve. He demonstrated a professional attitude and a remarkable ability to deliver requirements on time, even when faced with challenging situations. Jathurshan went beyond expectations by putting in extra effort to ensure project success.

RP
Rashmi PrabhathLead Software Engineer
"colleague

"Operated at a Software Engineer level, not an intern level."

Unlike a typical intern who performs only assigned tasks, he consistently operated at a Software Engineer level — taking full ownership, proactively identifying solutions, and delivering beyond expectations. Highly solution-oriented, always willing to explore alternatives and refine approaches.

IT
Irantha ThilakarathnaTechnical BA
"colleague

"Always asking the right questions."

From the first day onwards, Jathurshan impressed us with his technical skills. He was always asking the right questions, which helped him improve as well as contribute to the project. He was always responsible with his work and committed immensely.

HM
Haathim MunasSoftware Engineer
"colleague

"Stood out as a promising Software Engineer."

He consistently completed his tasks with accuracy and professionalism. Jathurshan collaborated exceptionally well with QA and Business Analyst teams, communicating clearly and contributing meaningful ideas.

SM
Shanilka MadushanSoftware Engineer
"colleague

"Hard work, accountability, strong problem-solving."

Impressed by his hard work, accountability, and strong problem-solving ability. He consistently took ownership, quickly understood requirements, and delivered high-quality work with a positive attitude.

SK
Sukumar KavishanSoftware Engineer
"colleague

"Quick to adapt, strong problem-solver, genuinely professional."

I had the pleasure of working with Jathurshan during his internship as a Software Engineering Intern at IronOne Technologies. From the beginning, he demonstrated a strong ability to quickly adapt to the project domain and understand complex requirements within a short period of time. He consistently showed excellent technical skills, problem-solving ability, and a willingness to contribute to critical project activities — proactive in understanding the project's critical path and always ready to support the team in delivering quality outcomes. What impressed me most was his dedication, fast learning capability, and the professionalism he maintained throughout his internship.

SV
Sanka VidushanQA Engineer

let's build
something.

I just wrapped my internship at IronOne and I'm looking for a full-time software engineering role — backend, devops / platform, ai/ml, or anywhere the lines blur. If you've got a system that should hold up, I'd love to talk.