┌──[SYS.USER_01]-[~]-[main]
└─> niroshan.exe
Niroshan 1.12.9 (tags/v3.12.6:a4a2d2b, dec 9 2001, 20:11:23) [MSC v.1940 64 bit (AMD64)] on linux
Building systems where mathematics meets markets.
>>> _
! [LOCATION] : Kandy, Sri Lanka
+ [STATUS] : CS Undergraduate · Final Year · University of Staffordshire
# [DOMAIN] : Quantitative Finance · Probabilistic ML · High-Performance Computing
- [LANGUAGES] : Python · C++ · JavaScript · Kotlin · C#Computer Science undergraduate specialising in Machine Learning for quantitative
finance and high-performance systems. I build probabilistic risk models, regime-
adaptive trading systems, and simulation engines — from scratch, in C++ and Python.
Seeking: ML Engineering · Quant Research · FinTech roles
drwxr-xr-x FINANCE_ML/
drwxr-xr-x SYSTEMS/
drwxr-xr-x FULLSTACK/
name : Bayesian Credit Risk Engine
stack : PyMC · MCMC · ArviZ · Streamlit
status : [████████████████████] DEPLOYED
year : 2026
link : https://github.com/Niroshan-k/Bayesian-Credit-Risk-Engine
desc : Probabilistic SME loan default model using Bayesian inference and MCMC
sampling. Produces full posterior distributions with 95% credible intervals.
Streamlit dashboard enables uncertainty-aware credit scoring and dynamic
interest rate pricing — aligned with Basel III IRB frameworks.name : Bayesian PSO Portfolio Optimizer
stack : C++ · PyBind11 · PyMC · MCMC
status : [████████████████████] DEPLOYED
year : 2026
link : https://github.com/Niroshan-k/Bayesian-Swarn-Portfolio-Optimizer
desc : High-performance particle swarm optimisation engine written in C++,
bridged to Python via PyBind11. Applies Bayesian inference to model
parameter uncertainty — producing robust, risk-adjusted portfolio weights
rather than a single fixed allocation.name : Regime-Adaptive Algorithmic Trader
stack : PCA · K-Means · XGBoost · GitHub Actions · Streamlit
status : [████████████████████] LIVE
year : 2026
link : https://github.com/Niroshan-k/Regime-Adaptive-Algorithmic-Trader
desc : Quantitative trading bot that detects market regimes via PCA and K-Means
clustering, then hot-swaps specialist XGBoost models per regime in real-time.
CI/CD via GitHub Actions. Live on Streamlit Cloud.name : FinBERT Stock Screener
stack : FinBERT · HuggingFace · yfinance
status : [████████████████████] DEPLOYED
year : 2026
link : https://github.com/Niroshan-k
desc : NLP pipeline using FinBERT to classify real-time financial news sentiment.
Dynamically scores and filters equities — turning news catalysts into
actionable trading signals.name : Ferrum — Event-Driven Monte Carlo Simulation Engine
stack : C++ · Python · FSM · pybind11
status : [█████████████░░░░░░░] IN PROGRESS
year : 2026
link : https://github.com/Niroshan-k/Event-Driven-Simulation-Framework
desc : Final year thesis project. Hybrid C++/Python engine for business risk
analysis using deterministic FSM-driven Monte Carlo simulation. Benchmarked
at significant speedups over pure Python and Backtrader. Architecture:
event-driven state machines over AI agent pattern recognition.name : Dimora Real Estate Platform
stack : MERN · Laravel · Kotlin · Firebase · MongoDB
status : [████████████████████] DEPLOYED
year : 2025
link : https://github.com/Niroshan-k/DimoraMernProject
desc : Full-stack real estate platform with Android client. Features secure user
verification, brute-force prevention, XSS protection, live admin dashboard,
and AI-assisted property search agents.[FINANCE_ML]
PyMC · MCMC · XGBoost · FinBERT · scikit-learn · PyTorch
Bayesian Methods · Portfolio Optimisation · Regime Detection · Credit Risk
[SYSTEMS]
C++ · Python · PyBind11 · Linux · Docker · Git · GitHub Actions
[LANGUAGES]
Python · C++ · JavaScript · Kotlin · C# · PHP
[BACKEND]
Node.js · Express.js · Firebase · React · MongoDB · SQL · Streamlit
[FROM_SCRATCH]
Neural Nets · Backpropagation · PSO · Neuroevolution · Gradient Descent[2026] JPMorgan Chase & Co. — Quant Research Job Simulation (Forage)
▸ Price data analysis and commodity storage contract pricing
▸ Credit risk analysis and FICO score bucketing
[2019–2021] Cadet Platoon Lance Corporal — Jinaraja College
▸ Led training coordination and team discipline
▸ Top 10 national placement — Hermann Loos & De Soysa Trophies
lakshan@niroshan:~$ uptime
currently: learning · building · shipping
interests: chess · buddhist philosophy · creative writing · cybersecurity
next: ML Engineering · Quant Finance · Open to internships
lakshan@niroshan:~/github$ █