About Me · Career Timeline · Flagship Projects · AtlasDLH-Bank · AtlasFlow · BRAINO · BoBo · Project Comparison · Data Engineering & BI Projects · Other Notable Projects · Tech Stack · GitHub Stats · Professional Summary · Contact
Data, AI/ML, and Cloud Engineer building enterprise-grade solutions across data lakehouses, AI systems, and cloud infrastructure — designed and shipped end-to-end, not prototyped.
- 🧠 Currently Building: Two independent, production-shaped banking data lakehouses — AtlasDLH-Bank (Docker-native) and AtlasFlow (cloud-native)
- 🎯 Latest Achievement: BRAINO — Enterprise AI Business Operations Manager, ✅ 100% COMPLETE, 30/30 parts, live in production
- 💼 Current Internship: Data Engineer Intern at Bank of Ceylon (BOC) — Head Office, contributing to an enterprise-scale Data Lakehouse implementation
- 💼 Working as: Data Engineer · AI/ML Engineer · Cloud Engineer
- 💼 Specializing in:
- Data Engineering: Apache Kafka, Debezium CDC, Apache Spark, Delta Lake, dbt, Apache Airflow, Trino, medallion architecture (Bronze → Silver → Gold)
- Cloud & DevOps: Kubernetes, Docker, Terraform (IaC), AWS, BigQuery, GitHub Actions CI/CD, HPA autoscaling, Prometheus/Grafana monitoring
- AI/ML: RAG Systems, LangGraph multi-agent orchestration, LLM Integration, Vector Databases, MLflow model registries, PyTorch, Financial AI
- Security & Compliance: PII tokenization, Fernet encryption, JWT RBAC, audited AML-grade detokenization workflows
- Full-Stack: MERN, React, Next.js, Flask, FastAPI, Node.js, React Native/Expo
- Mobile: Android (Kotlin/Java), Material Design, Responsive UI
- UI/UX: Figma, User-Centered Design, Prototyping
- Data Warehousing & BI: Star Schema, SSAS, Power BI, Tableau, Statistical Analytics with R
- 🔭 Tech Arsenal: Python, TypeScript, JavaScript, Kotlin, Java, SQL, R, PostgreSQL, MongoDB, Kafka, Spark, Kubernetes, Docker, Terraform, AWS
- 💡 Portfolio: 13+ completed projects spanning enterprise data platforms, AI/ML systems, cloud infrastructure, web, mobile, and design
- 🌐 Live / Deployed Projects: BRAINO (Production), AtlasDLH-Bank, AtlasFlow, SafeZone DMS, Travellers Platform, SkyWay, TripMate, Portfolio
- 👯 Open to: Data Engineer, AI/ML Engineer, Cloud Engineer, Data Platform Engineer roles
- 💬 Ask me about: Medallion architectures, PII tokenization, RAG systems, LangGraph agents, Kubernetes deployments, CI/CD pipelines, MERN stack, system architecture
- 📫 Contact: pasindialawatta@gmail.com
- ⚡ Fun fact: Built an enterprise AI system with 34,500+ lines of code in 8 weeks — and then went on to build two full banking data lakehouses.
| Strength | Evidence |
|---|---|
| Enterprise architecture thinking | Two independent, full medallion-architecture banking lakehouses (Docker-native + cloud-native) with PII-gated pipelines |
| AI systems that ship | 3 production-shaped AI copilots (RAG + LLM) across different stacks — Ollama, Groq, LangGraph |
| Security-first engineering | Token vaults, Fernet encryption, JWT RBAC, and audit-logged detokenization workflows modeled on real AML compliance flows |
| End-to-end ownership | From Terraform/IaC and CI/CD to Kubernetes deployment to the dashboard a business user clicks |
| Cloud & platform engineering | Docker Compose (14+ services), Kubernetes with HPA autoscaling, Terraform-provisioned BigQuery, GitHub Actions CI/CD, Prometheus/Grafana monitoring |
| Documentation discipline | Every flagship project ships with phase checklists, architecture diagrams, and verification scripts |
2026 ── Data Engineer Intern @ Bank of Ceylon (BOC), Head Office
│ └─ Enterprise Data Lakehouse implementation (production banking scale)
│
├── AtlasDLH-Bank ── 11/11 phases complete
│ Docker-native medallion lakehouse · PII tokenization · AI copilot · compliance API
│
├── AtlasFlow ── 5/5 steps complete
│ Cloud-native rebuild · dbt + Terraform + Kubernetes + LangGraph agents
│
├── BRAINO ── 30/30 parts complete · LIVE in production
│ Enterprise AI Business Operations Manager · 34,500+ lines · Railway + Vercel
│
├── BoBo ── Phases 1, 2, 5 complete · 3, 4, 6 in progress
│ Cross-border retail platform · monorepo · mobile + web + API
│
└── Data Warehousing & BI, Full-Stack, Mobile, and UI/UX projects (10+)
NYHospitalDW · Financial Tracker · Doctor-Patient Analysis · SafeZone DMS ·
Travellers Platform · SkyWay · TripMate · Sky Mountain · Wellness Tracker ·
EduFlex · Portfolio · To-Do List
Four enterprise-scale systems — architected, built, tested, and shipped end-to-end.
A fully containerized banking data lakehouse that ingests raw customer and transaction data from core-banking systems, moves it through an industry-standard Bronze → Silver → Gold medallion pipeline, and delivers executive dashboards, ML fraud/churn scores, natural-language analytics, and a JWT-audited compliance API — all while keeping personally identifiable information completely out of the analytics and ML path.
| Outcome | How the platform delivers it |
|---|---|
| Single source of truth for KPIs | Gold layer + Trino SQL + Superset dashboards |
| Fraud & churn intelligence | ML models trained on tokenized features, scored daily |
| Self-service analytics | AI Copilot answers questions in plain English |
| Regulatory privacy | PII encrypted on-prem; analytics use tokens only |
| AML investigations | Compliance officers detokenize with full audit trail |
Phase-by-phase breakdown (click to expand)
| Phase | Delivered | Key artifacts |
|---|---|---|
| 1 — Core Platform | ✅ | Docker Compose stack: PostgreSQL, MySQL, Kafka/Debezium, MinIO, Spark, Airflow, Trino, Superset, Prometheus/Grafana/Loki, MLflow |
| 2 — Ingestion & CDC | ✅ | 500 synthetic customers, ~1,000 accounts, 20,000+ transactions; Debezium CDC on Postgres + MySQL; 9 Kafka topics; Bronze Delta ingestion |
| 3 — PII Gate & Token Vault | ✅ | PII classification gate before any Bronze write; Fernet-encrypted pii_store; deterministic TOK_… tokens; atlas-bronze-pii bucket |
| 4 — Silver Layer | ✅ | 6 tokenized Silver Delta tables (customers_clean, accounts_clean, transactions_clean, loans_clean, atm_clean, mobile_clean) with PII guard |
| 5 — Gold Layer | ✅ | 8 Gold tables: customer_360, daily_transactions, loan_portfolio, fraud_summary, branch_performance, executive_kpis, ml_features_fraud, ml_features_churn |
| 6 — Data Quality & Governance | ✅ | Great Expectations suites (PII guard, positive amounts, NPA% bounds); OpenMetadata catalog/lineage |
| 7 — Analytics & BI | ✅ | 4 Superset dashboards (Executive, Customer, Fraud, Branch) over 16 Trino-backed datasets |
| 8 — Machine Learning | ✅ | Fraud detection, churn prediction, credit risk models — MLflow registry, weekly retrain, daily batch scoring |
| 9 — AI Copilot | ✅ | FastAPI + RAG (ChromaDB) + Ollama — natural language → Trino SQL → plain-English answer |
| 10 — Detokenization & Compliance API | ✅ | JWT RBAC (Admin/Compliance Officer/Auditor…), audit-logged AML identity resolution, 7-year audit retention |
| 11 — Atlas Command Center | ✅ | Unified executive portal — 17 dashboards, Layer Health, Customer Portfolio, Report Center, Identity Resolution UI, ML Trainer |
PII Gate & Vault Addressing (click to expand)
Source DBs (PostgreSQL + MySQL)
│
│ PII Classification Gate (Phase 3) — runs BEFORE any lakehouse write
├──────────────────┬─────────────────────┐
▼ ▼ ▼
pii_store + atlas-bronze token_vault
atlas-bronze-pii (non-PII + token) (token ↔ customer_id)
(on-prem PG) │ (on-prem PG)
│ ▼
│ atlas-silver (6 clean tables, tokens only)
│ │
│ ▼
│ atlas-gold (KPIs, customer_360, ML features)
│ │
│ ┌───────────┼───────────┬─────────────┬──────────────┐
│ ▼ ▼ ▼ ▼ ▼
│ Trino Superset MLflow AI Copilot Command Center
│ │
│ click token / PII │
└──── Detokenization API ◄── Identity Resolution UI (audited, JWT RBAC)
RBAC matrix:
| Role | Bronze PII | Silver/Gold | Detokenize | Audit Log |
|---|---|---|---|---|
| ADMIN | ✅ | ✅ | ✅ | ✅ |
| DATA_ENGINEER | ❌ | ✅ write | ❌ | ❌ |
| DATA_ANALYST | ❌ | ✅ read | ❌ | ❌ |
| COMPLIANCE_OFFICER | ❌ | ✅ read | ✅ | ✅ |
| AUDITOR | ❌ | ✅ read | ✅ | ✅ read-only |
A single executive URL replacing Superset + Airflow + MLflow + Security Swagger: Mission Control (stale-while-revalidate KPI cache), Medallion Layer Health, Customer Portfolio (branch-desk lookup with loan-readiness checklist), Report Center (PDF/Excel/CSV exports), Identity Resolution, and an ML Trainer with MLflow-backed versioning.
Tech: 
The same banking domain as AtlasDLH-Bank, rebuilt the way a modern cloud-native data team ships it: Infrastructure-as-Code, CI/CD-gated transforms, containerized model serving, and specialized AI agents — free-tier only (BigQuery Sandbox, Colab T4, minikube).
┌──────────────────────────────────────────────────────────────────────────┐
│ SOURCE — AtlasDLH-Bank Gold exports → AtlasFlow/seeds/*.csv │
└────────────────────────────────┬─────────────────────────────────────────┘
▼
┌──────────────────────────────────────────────────────────────────────────┐
│ STEP 1 — FOUNDATION │
│ Terraform → BigQuery dataset (cloud) · dbt: seeds → staging → marts │
│ Marts: mrt_fraud_features · mrt_churn_features · mrt_executive_summary │
└───┬───────────────────────────────┬──────────────────────────┬───────────┘
▼ ▼ ▼
┌─────────────────────┐ ┌─────────────────────┐ Portal + SQL agent
│ STEP 2 — DATAOPS │ │ STEP 3 — ML │ (reads same Gold)
│ GitHub Actions CI │ │ PyTorch fraud/churn │
│ Airflow scheduling │ │ models → MLflow │
└─────────────────────┘ └──────────┬──────────┘
▼
┌─────────────────────┐
│ STEP 4 — SERVING │
│ FastAPI → minikube │
│ HPA + Prometheus │
└─────────────────────┘
STEP 5 — COPILOT + PORTAL
Portal (Streamlit :8501) ← DuckDB Gold marts
SQL agent + RAG agent (ChromaDB policies) + Compliance agent
LangGraph orchestration ← Ollama llama3.2 :11434 ← Copilot API :8097
All 5 steps in detail (click to expand)
| Step | What it does | Key components |
|---|---|---|
| 1 — Foundation | Seeds → dbt three-layer project (staging → intermediate → marts) on DuckDB (local) + BigQuery (cloud, via Terraform) | dbt_project/, terraform/, schema tests |
| 2 — DataOps | Keeps Step 1 correct & repeatable | GitHub Actions (dbt CI, lint, prod deploy), Airflow DAG (seed → run → test) on port 8091 |
| 3 — ML | Learns fraud/churn patterns from Gold features | PyTorch models, MLflow tracking (port 5002), optional Llama QLoRA fine-tune on Colab free T4 |
| 4 — Serving | Serves the fraud model as a production-shaped API | FastAPI /predict, Docker → minikube, HPA (2–5 pods), Prometheus + Grafana (30300) |
| 5 — Copilot + Portal | Human-facing AI + dashboards | LangGraph intent router → SQL agent / RAG agent → Compliance agent; Streamlit Portal (5 pages) |
| Page | Data source | Purpose |
|---|---|---|
| Executive Overview | mrt_executive_summary |
Leadership KPIs |
| Fraud Risk Desk | mrt_fraud_features |
Risk mix + review queue |
| Churn Watch | mrt_churn_features |
Inactivity / outreach list |
| Banking Copilot | LangGraph API | SQL + policy Q&A |
| Platform Control | Health probes | Prove all 5 steps are live |
Tech: 
A commercial-grade AI system combining document intelligence, conversational AI, financial intelligence, compliance monitoring, proactive monitoring, weakness detection, and interactive problem-solving — FULLY DEPLOYED in production on Railway (backend) and Vercel (frontend).
|
|
|
|
|
|
Backend: FastAPI (Python) | Groq API (Llama 3.3 70B) | PostgreSQL | ChromaDB | Redis | Celery Frontend: Next.js 14 | TypeScript | Tailwind CSS | Recharts | Radix UI AI/ML: RAG Pipeline | sentence-transformers | Vector Embeddings (384-dim) Payments: Stripe Integration | Subscription Management Email: SendGrid Integration | Automated Notifications DevOps: Railway | Vercel | Docker | Git
📋 All 30 Completed Parts (click to expand)
- ✅ Secure Authentication (JWT + bcrypt)
- ✅ File Upload System (ClamAV virus scanning)
- ✅ Document Parsing (PDF, DOCX, CSV, MD, TXT)
- ✅ Smart Chunking (semantic boundaries)
- ✅ Background Tasks (Celery + Redis)
- ✅ Local Embeddings (sentence-transformers)
- ✅ Vector Search (ChromaDB <100ms)
- ✅ RAG Pipeline (Groq + Llama 3.3 70B)
- ✅ Citation System (exact sources)
- ✅ Conversation Memory (10-msg context)
- ✅ Security Hardening (7 headers + AES-128)
- ✅ High-Confidence Mode (industry validation)
- ✅ Feedback System (corrections + metrics)
- ✅ Analytics Dashboard (usage stats)
- ✅ Team Management (workspaces + RLS)
- ✅ Chat Interface (real-time streaming)
- ✅ Document Stats (word/page count)
- ✅ Document Management (CRUD + metadata)
- ✅ Slack Integration (slash commands)
- ✅ Production Deploy (Railway + Vercel)
- ✅ Financial Intelligence (transactions + budgets)
- ✅ Compliance Monitoring (rules + alerts)
- ✅ Operational Analytics (metrics + KPIs)
- ✅ Smart Alerts & Notifications (SendGrid)
- ✅ Report Generation (PDF reports)
- ✅ Braino AI (business operations advisor)
- ✅ Proactive Monitoring (24/7 health checks)
- ✅ Weakness Detection (7 categories)
- ✅ Recommendation Engine (AI-powered)
- ✅ Interactive Problem Solving (guided workflows)
Lines of Code: 34,500+ | API Endpoints: 120+ | Database Tables: 30+
Search Speed: <100ms | RAG Response: <200ms | Dashboard Load: <500ms
Citation Accuracy: 98% | Follow-up Detection: 95% | Workspace Isolation: 100%
Anomaly Detection: 85%+ | Health Monitoring: 24/7 | Uptime: 99.9%
Test Suites: 31/31 passing | Documentation: 18,000+ lines
| Metric | Value | Status |
|---|---|---|
| Total Parts | 30/30 | ✅ 100% Complete |
| Total Phases | 7/7 | ✅ All Done |
| Development Time | 8 weeks | ✅ Finished |
| Production Status | Live | ✅ Deployed |
| API Endpoints | 120+ | ✅ All Working |
| Test Suites | 31 | ✅ All Passing |
| Documentation | 18,000+ lines | ✅ Complete |
Tech: 
A cross-border retail platform letting customers in Sri Lanka buy authentic imported products from Australia, Japan, USA, and Canada — shipped as a mobile app (Expo/React Native, Google Play + App Store), a responsive web storefront (Next.js), and a fully-tested REST API (Express + Prisma) in a pnpm monorepo.
Full route table (click to expand)
| Domain | Routes | RBAC |
|---|---|---|
| Auth | register, get/update/delete profile | All roles |
| Products | list, get, create, update, delete, reviews (get/post) | Admin/Supplier write, public read |
| Orders | list, get, create, update status, cancel, invoice | Customer create, Admin/Delivery manage |
| Inventory | list, low-stock, update stock, bulk update | Admin/Supplier |
| Shipments | get, create, update status, add tracking event | Admin/Delivery |
| Payments | PayPal create/capture, PayHere webhook, live FX rate | Customer / public webhook |
| Users | list, get, update, list orders, addresses CRUD | Admin/self |
| Uploads | single/multi product image upload, delete image | Admin/Supplier |
|
Commerce
|
Platform Engineering
|
BoBo/
├── apps/
│ ├── api/ — Express + Prisma + Socket.io (✅ Phase 2 complete — 38 routes)
│ ├── web/ — Next.js 14 storefront (Phase 4 — 75% complete)
│ └── mobile/ — Expo Router + NativeWind (Phase 3 — in progress)
└── packages/ — @bobo/types · @bobo/config · @bobo/ui (shared across apps)
PENDING → CONFIRMED → PROCESSING → SHIPPED → OUT_FOR_DELIVERY → DELIVERED → REFUNDED
│
└──────────────────────── CANCELLED (from PENDING) ─────────────────
Tech: 
| 🏦 AtlasDLH-Bank | ☁️ AtlasFlow | 🧠 BRAINO | 🛍️ BoBo | |
|---|---|---|---|---|
| Domain | Banking data lakehouse | Cloud-native AI/ML platform | Enterprise AI ops manager | Cross-border retail |
| Status | 11/11 phases ✅ | 5/5 steps ✅ | 30/30 parts ✅ live | Phases 1,2,5 ✅ |
| Core engine | Spark + Delta Lake + Trino | dbt + DuckDB/BigQuery | FastAPI + RAG | Express + Prisma |
| AI layer | RAG + Ollama copilot | LangGraph multi-agent | RAG (Groq Llama 3.3 70B) | — |
| Orchestration | Airflow (6 DAGs) | Airflow + GitHub Actions | Celery + Redis | — |
| Serving | Docker Compose (14 services) | Kubernetes (HPA) | Railway + Vercel | Railway/Vercel-ready |
| Security model | PII token vault + JWT RBAC + audit | Compliance agent (rule-based) | JWT + AES-128 + RLS | Clerk JWT + RBAC |
| Frontend | Atlas Command Center (17 dashboards) | Streamlit Portal (5 pages) | Next.js dashboard | Next.js + Expo |
| Deployment cost | $0 (self-hosted OSS) | $0 (free tiers) | Live in production | Local / staging |
Full Star Schema data warehouse and BI stack built on real-world New York State hospital admission data, covering everything from dimensional modeling to cloud-published BI reports - developed for IT3021: Data Warehousing & Business Intelligence at SLIIT.
🔍 Key Features & Tech Details (click to expand)
- 🗂️ Data Warehouse: NYHospitalDW with a central FactAdmission fact table and 6 dimension tables (DimDate, DimHospital, DimPatient, DimDiagnosis, DimProcedure, DimPayment)
- 🧊 OLAP Cube: SSAS multidimensional cube with dimension hierarchies and measure groups, deployed via an XMLA script through SSMS after a Visual Studio deployment issue
- 📊 OLAP Analysis: Roll-up, drill-down, slice, dice, and pivot operations on live cube data using Excel PivotTables for business-friendly exploration
- 📈 Power BI Reports: 4 interactive reports (Matrix, Slicers, Drill-Down, Drill-Through) published to Power BI Service for cloud-based access and sharing
- Database: SQL Server (SSMS), T-SQL
- Analytics: SQL Server Analysis Services (SSAS), XMLA
- BI & Reporting: Power BI Service, Excel PivotTables
- Concepts: Star Schema, Dimensional Modeling, ETL Pipeline Design
Repository: GitHub - NYHospitalDW & BI Solution
Full-stack personal finance tracker built as a Data Engineering internship project, combining a Laravel REST API with interactive grids, pivot reporting, and export pipelines.
🔍 Key Features & Tech Details (click to expand)
- 🔐 REST API: Laravel 11 with Sanctum authentication
- 📊 Interactive Data Grids: Tabulator with inline editing, multi-column filters, and pagination
- 📅 Monthly Pivot Tables: Category-wise spending breakdowns
- 🎯 Budget vs. Actual Tracker: Colour-coded status indicators
- 📈 Trend Charts: ApexCharts for income vs. expense visualization
- 📑 Export Layer: Excel and PDF reporting
- Backend: Laravel 11, Sanctum, MySQL
- Frontend: Tabulator, ApexCharts, Tailwind CSS
- Concepts: ETL pipelines, fact/dimension data modeling, aggregation & reporting layers
Repository: GitHub - Financial Tracker
End-to-end statistical and predictive analytics project exploring whether doctor-patient communication quality predicts a patient's ability to understand their next steps, using healthcare survey data.
🔍 Key Features & Tech Details (click to expand)
- 🧹 Data cleaning and preprocessing
- 📊 Exploratory and descriptive data analysis
- 📈 Data visualization with ggplot2
- 🔗 Correlation analysis
- 📉 Linear regression modeling
- ✅ Model diagnostics and validation
- 🤖 Predictive analytics: Linear Regression & K-Nearest Neighbors (KNN)
- Language: R
- Visualization: ggplot2
- Concepts: Statistical modeling, predictive analytics, healthcare data analysis
Repository: GitHub - Doctor-Patient Communication Analysis
Full-stack disaster response coordination system with real-time shelter management, victim tracking, donation coordination, and interactive mapping.
🔍 Key Features & Tech Details (click to expand)
- 🗺️ Interactive Maps: Real-time shelter locations with Mapbox, user geolocation, evacuation routes
- 🏠 Shelter System: Find nearest shelters, geospatial queries, CRUD operations
- 👥 Victim Management: Registration, disaster reporting, aid requests, claim tracking
- 💝 Donation Dashboard: Track donations, inventory, volunteers, distribution plans
- 📊 Admin & DMO Dashboards: Alerts, reports, deployments, response team management
- 🌤️ Weather Broadcast: Automated weather alerts for Sri Lanka via cron jobs
- Backend: Node.js | Express.js | MongoDB | Mongoose
- Frontend: React | Material-UI | Chakra UI | Mapbox
- Architecture: RESTful API | MVC Pattern | Modular CSS Scoping
- Features: Email notifications | Session management | Geospatial queries
- Modular, conflict-free CSS architecture
- Real-time weather updates
- Geospatial search with MongoDB queries
- Scalable RESTful API design
Repository: GitHub - SafeZone DMS
Comprehensive travel booking platform with flight bookings, hotel reservations, visa services, and secure payment processing.
🔍 Key Features & Tech Details (click to expand)
- 🔐 User Management: Registration, authentication, profile management
- 🌍 Destination Discovery: Browse and search travel destinations
✈️ Flight Booking: Search and book flights with multiple airlines- 🏨 Hotel Reservations: Find and book accommodations
- 🚗 Transportation: Ground transportation booking
- 🛂 Visa Services: Visa requirement checking and application
- 💳 Payment Processing: Secure payment handling with Stripe/PayPal
- ⭐ Reviews & Ratings: User review system
- 👨💼 Admin Dashboard: Comprehensive management panel
- Backend: Flask (Python) | SQLAlchemy | PostgreSQL/SQLite
- Frontend: React 18 | Tailwind CSS | React Query | Context API
- Authentication: JWT tokens
- API: RESTful design
- Payments: Stripe & PayPal integration
- Email: Flask-Mail for notifications
- Forms: React Hook Form
- Routing: React Router v6
- Notifications: React Hot Toast
- Animations: Framer Motion
- Clean separation of concerns
- Component-based architecture
- State management with React Query + Context
- Responsive design for all devices
Repository: GitHub - Travellers Platform
Modern Android travel/flight booking app with Material Design, fragments architecture, and comprehensive form validation.
🔍 Key Features & Tech Details (click to expand)
- 📱 Flight Booking: Complete flight search and booking workflow
- 📋 Ticket Management: View, manage, and track bookings
- 🎫 Digital Boarding Pass: Mobile boarding pass generation
- 🔍 Flight Search: Advanced search with filters
- 📊 Booking History: Track past and upcoming flights
- 🌙 Dark/Light Theme: User preference support
- Language: Kotlin
- UI: Material Design principles
- Architecture: Fragment-centric modularity
- Navigation: Manual fragment transactions with Bundle arguments
- Components: Reusable card components
- Forms: Advanced validation mechanics (email, password complexity)
- Data Generation: Dynamic time offsets, ticket samples
- Resources: Centralized strings, colors, dimensions, arrays
- Responsive: Works on smartphones and tablets
- Fragment lifecycle management
- Manual navigation with back stack control
- Time mapping and autocomplete
- Grid & list implementations (RecyclerView)
- Weighted columns for responsive layouts
- Data binding without heavy frameworks
- Lightweight with Shared Preferences
- Smooth responsive experience
- Material Design animations
- Optimized for battery usage
Documentation: Comprehensive viva/oral defense documentation with 28 sections covering architecture, patterns, and best practices
Personal travel planner prototype with clean, modern, user-friendly experience designed in Figma.
🔍 Key Features & Design Details (click to expand)
- 🌍 Discover Best Countries: Browse and explore top travel destinations
- 🏨 Hotel Booking: Explore and book hotels for your stay
✈️ Air Tickets: Book flights with ease- 🚂 Transport Booking: Reserve train and public transport
- 🎫 Digital Boarding Pass: View your boarding pass digitally
- 📱 Instant Notifications: Get alerts when bookings are confirmed
- 👤 Traveler Profile: Manage your profile with ease
- Clean & Modern: Minimalist design aesthetic
- User-Friendly: Intuitive navigation and interactions
- Freedom & Control: Give travelers complete autonomy
- Convenience: Everything accessible at fingertips
- Professional: Industry-standard UI/UX practices
- User research and persona development
- Wireframing and prototyping
- High-fidelity mockups
- Interactive prototype with Figma
- User flow optimization
Goal: Eliminate reliance on travel agencies by empowering users with a comprehensive self-service platform
Online Property Management System built as SLIIT OOP group project with Java, HTML, CSS, and MySQL.
🔍 Key Features & Tech Details (click to expand)
- 🔐 User Authentication: Login system
- 👤 User Profile: Create and update user profiles
- 👨💼 Admin Dashboard: User management (read, update, delete)
- 🔧 Servlets: All user-related servlet pages
- userDeleteServlet.java
- userInsertServlet.java
- userLoginServlet.java
- userProfileServlet.java
- userProfileUpdateServlet.java
- userReadServlet.java
- userUpdateServlet.java
- 📄 Views: JSP files (userLogin, userProfile, userRegistration, userUpdate, userDisplay)
- 🎨 Frontend: HTML pages, CSS styling
- 📝 Logging: Logger utility for all user servlets
- 🔧 Constants: User details configuration
- Backend: Java | Eclipse IDE
- Frontend: HTML | CSS | JavaScript
- Database: MySQL
- Architecture: MVC (Model-View-Controller)
- Design Patterns: Singleton (DB connection), Custom Logger Utility
- Features: Session Management | Admin Dashboard
- OOP principles in practice
- MVC architecture implementation
- Java-based full-stack development
- Scalable design patterns
- Team collaboration in real-world projects
Repository: GitHub - Sky Mountain
Professional portfolio website showcasing projects, apps, websites, and UI/UX designs.
🔍 Key Features & Tech Details (click to expand)
- 📂 Project Showcase: Display completed applications, websites, and UI/UX designs with detailed descriptions and visuals
- 🎓 Professional Qualifications: Education, certifications, and technical skills
- 📄 CV Download: Direct download of updated resume for employers/recruiters
- 📞 Contact Section: Easy-to-use form and direct contact links
- ✨ Professional UI: Clean, minimal, modern design reflecting industry standards
- Frontend: HTML | CSS | JavaScript | React
- Design: Figma for mockups and prototyping
- Backend: Java (for dynamic features)
- Responsive: Mobile-first design approach
- Minimal and professional aesthetic
- Fast loading times
- Easy navigation
- Clear call-to-actions
- Accessibility-focused
Purpose: Central hub for potential employers, collaborators, and clients to explore my work
Lightweight task management tool with CRUD operations and persistent local storage.
🔍 Key Features & Tech Details (click to expand)
- ➕ Add Tasks: Instantly add new tasks
- 🖊️ Edit Tasks: Modify existing tasks
- 🗑️ Delete Tasks: Remove completed or unwanted tasks
- ✅ Mark as Completed: Track completion status
- 💾 Persistent Storage: Data saved locally (localStorage)
- 📱 Responsive Design: Clean UI for distraction-free experience
- Backend: Python
- Frontend: HTML | CSS | JavaScript
- Storage: Browser localStorage for data persistence
- Design: Clean and minimal interface
- Daily schedule management
- Personal reminders
- Goal tracking
- Task prioritization
Repository: GitHub - To-Do List
Online Teacher Training platform developed as 1st year group project at SLIIT.
🔍 Key Features & Tech Details (click to expand)
- 👨🏫 Teacher Training: Comprehensive online training platform
- 📝 Course Management: Create and manage training courses
- 👥 User Management: Teacher and admin accounts
- 📊 Progress Tracking: Monitor learning progress
- 💬 Simple & User-Friendly: Easy navigation and clean interface
- Frontend: HTML | CSS | JavaScript
- Backend: PHP
- Database: SQL
- Architecture: Traditional server-side rendering
University first year, 2nd semester group project demonstrating foundational web development skills
Repository: GitHub - EduFlex
Comprehensive wellness app for tracking daily habits, moods, and hydration - developed for SLIIT Mobile App Development Lab Exam.
🔍 Key Features & Tech Details (click to expand)
- 📋 Daily Habit Tracker: Create, update, delete, monitor habits with visual progress tracking
- 😊 Mood Journal: Emoji selector with calendar/list views for reflection and sharing
- 💧 Hydration Reminder System: Alarm Manager and Work Manager for timely drinking reminders
- 🏠 Home Screen Widget: Quick access to progress updates directly from home screen
- 📊 Progress Tracking: Visual progress bars and completion statistics
- 💾 Data Persistence: Efficient management through Shared Preferences
- Language: Kotlin
- UI: Material Design principles
- Architecture: Fragment-based navigation
- Components: Bottom navigation, custom cards, calendars
- Data: Shared Preferences for lightweight storage
- Notifications: Alarm Manager for reminders
- Background: Work Manager for reliable scheduling
- Design: 60/30/10 color scheme (black/white/accent)
- Custom mood tracking with trend analysis (7-day chart)
- Configurable reminder intervals (customizable minutes)
- Data export/import functionality (JSON format)
- Sample data generation for testing
- Clear all data with confirmation
- Widget for home screen quick access
- Calendar view for mood history
Total Score: 10/10
- Functionality: 3/3 (Core and bonus features)
- UI/UX Design: 2/2 (Clean and intuitive)
- Code Quality: 2/2 (Well-organized and documented)
- Advanced Features: 3/3 (SharedPreferences + widget)
- Clean separation of concerns
- Reusable UI components
- Centralized resource management (strings.xml, colors.xml)
- Dark theme support
- Responsive layouts for phones and tablets
Documentation: Complete lab exam report with screenshots and technical documentation
Generated automatically every 24 hours (and on every push to
main) by thePlatane/snkGitHub Action in.github/workflows/snake.yml— no manual updates needed.
|
|
Working as a Data / AI / ML / Cloud Engineer, seeking opportunities in Data Engineering, AI/ML Engineering, Cloud Engineering, or Data Platform Architecture roles where I can leverage my expertise in building production-grade data platforms, enterprise AI solutions, and cloud-native infrastructure.
- ✅ Proven ability to build complex enterprise systems from scratch (two full banking data lakehouses, BRAINO - 34,500+ lines)
- ✅ Strong foundation in data engineering — Kafka, Spark, Delta Lake, dbt, Airflow, Trino
- ✅ Strong foundation in AI/ML, RAG, LangGraph agents, and LLM integration
- ✅ Full-stack expertise across multiple frameworks (MERN, Flask, FastAPI, Next.js)
- ✅ Mobile development experience (Android/Kotlin, React Native/Expo)
- ✅ UI/UX design capabilities (Figma)
- ✅ Production deployment experience (Railway, Vercel, Kubernetes, Docker)
- ✅ Team collaboration and project leadership
- ✅ Passion for clean code, documentation, and best practices
- ✅ 100% project completion rate
⭐ From pasindi15 — Building enterprise-grade data & AI systems, one pipeline at a time ⭐