Faizan Risk Management AI & Machine Learning Mobile Apps Tax Planning Business Solutions Contact
Chartered Accounting — ACCA · Pakistan

Numbers.
Risk.
Everything in between.

Chartered Accountant Affiliate by qualification. Risk analyst, ML builder, and app developer by compulsion. If it involves capital, code, or complexity — it's usually on my screen.

Chartered Accounting Risk Management Financial Modelling Python & ML UK Tax DeFi & Crypto React / Next.js Geopolitics
Scroll to explore
Chartered
Accountant
ACCA Affiliate
7+
Domains Covered
Always Learning

Chartered Accountant Affiliate (ACCA) based in Pakistan. I qualified through one of the most rigorous accountancy programmes in the world — and then immediately started breaking the mould. Audit, UK tax, financial planning are my professional foundation. But where my mind actually lives is in the intersection of risk, capital markets, and the systems that move money at scale.

I've built that out with a parallel track in software development — shipping mobile apps, building AI/ML models, and writing Python that models what most analysts can only spreadsheet. Not a side interest. A core part of how I operate.

ACCA Affiliate Audit UK Tax Financial Planning Pakistan
⚖️

Risk & Finance

Liquidation heatmaps, order flow, macro gauges, stress testing — the full stack of risk intelligence.

🤖

AI & ML Models

Random Forest classifiers, LSTM networks, and custom indicators built in Python for trading edge.

📱

Mobile Development

React Native, Next.js, Node.js — professional-grade apps built alongside the CA qualification.

🌐

Macro Interests

Fed PCE, CPI, geopolitics, DeFi regulation — tracking the systems that determine market structure.

Where the real
work happens.

Liquidation heatmaps. Order book depth. Macro gauges. This is the layer of analysis most people don't see — real-time risk intelligence applied to capital allocation decisions. Summarised views on USD, Gold, BTC, and ETH form part of my regular analytical workflow.
Liquidation Heatmap Real-time cluster analysis
$97,420
▲ 2.34%
BTC / USD Candlestick · 4H
Order Book Depth Bid / Ask spread analysis
$2,318
▲ 0.82%
Gold / USD · 1D
$3,241
ETH / USD
▲ 3.11%
Multi-Asset Overview XAU · ETH · macro correlation
⚠️ Disclaimer: The analysis and tools presented here are for informational and educational purposes only. I am not a financial advisor and nothing on this site constitutes financial advice. Always conduct your own research before making any investment decisions.

Sophisticated models.
Real-world edge.

Beyond spreadsheets. I build and deploy quantitative risk tools using Python — the same class of models used by CFA-level analysts and institutional desks.

faizan_risk_suite — Python 3.12 — Samsung Odyssey G9
value_at_risk.py — Monte Carlo VaR
# Monte Carlo Value-at-Risk
import numpy as np
from scipy.stats import norm

class MonteCarloVaR:
  def __init__(self, portfolio, sims=100_000):
    self.portfolio = portfolio
    self.sims = sims

  def compute(self, confidence=0.99):
    returns = np.random.multivariate_normal(
      self.portfolio.mu,
      self.portfolio.cov, self.sims)
    pnl = returns @ self.portfolio.weights
    var = np.percentile(pnl,
      (1-confidence)*100)
    es = pnl[pnl <= var].mean()
    return {"VaR_99": var,
         "ES_99": es}
stress_test.py — Scenario Engine
# Historical & Hypothetical Scenarios
from dataclasses import dataclass

SCENARIOS = {
  "GFC_2008": {"equity": -0.52,
    "credit": +0.45,"fx": -0.18},
  "COVID_2020": {"equity": -0.34,
    "vix": +0.300,"oil": -0.67},
  "FED_TAPER": {"rate_shock": +2.0,
    "duration_loss": -0.16},
}

def run_scenario(portfolio, scenario):
  shocks = SCENARIOS[scenario]
  pnl = portfolio.apply_shocks(shocks)
  return RiskReport(pnl, scenario)
risk_rating.py — Composite Scoring
# Composite Risk Rating Framework
class RiskScorer:
  WEIGHTS = {
    "market_vol": 0.30,
    "credit_risk": 0.25,
    "liquidity": 0.20,
    "concentration": 0.15,
    "macro_regime": 0.10
  }

  def score(self, data) -> dict:
    scores = {k: self._norm(data[k])
      for k in self.WEIGHTS}
    composite = sum(scores[k] *
      self.WEIGHTS[k] for k in scores)
    return {"composite": composite,
      "grade": self._grade(composite)}
terminal — Risk Dashboard Output
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
▶ RISK SUITE v2.4 — PORTFOLIO ANALYSIS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Portfolio: MULTI_ASSET_01
Positions: 14
AUM: $4.2M

Monte Carlo VaR (99%, 1D)
VaR: -$82,440
ES (CVaR): -$104,210

Stress Testing
GFC 2008: -$682,100
COVID 2020: -$441,600
Fed Taper: -$198,300

Risk Rating
Composite: 6.2 / 10
Grade: MODERATE-HIGH

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Monte Carlo VaR
100K simulation paths for portfolio-level Value-at-Risk and Expected Shortfall at 95/99% confidence.
Stress Testing
GFC, COVID, and Fed taper scenarios applied via historical shock matrices to capture tail-risk exposure.
Risk Rating Engine
Composite scoring across market vol, credit, liquidity, concentration, and macro regime — weighted multi-factor output.
Greeks & Duration
Option greeks decomposition and fixed-income duration/convexity modelling for interest rate sensitivity analysis.
USD
DXY 104.2
▲ 0.31%
US Dollar Index
XAU
$2,318
▲ 0.82%
Gold / USD · Safe Haven
BTC
$97,420
▲ 2.34%
Bitcoin · Digital Gold
ETH
$3,241
▲ 3.11%
Ethereum · DeFi Layer

Built alongside
the CA. Seriously.

While most accountants were mastering Excel shortcuts, I was learning React and shipping mobile apps. Full-stack development isn't a hobby — it's a parallel professional skill that shapes how I solve problems. Productivity apps, fintech tools, portfolio dashboards — all built professionally from scratch.

Tech Stack
React Native Next.js 14 Node.js TypeScript Python PostgreSQL Tailwind CSS Expo Supabase Vercel
App.tsx api/risk.ts portfolio.py
import { useState, useEffect } from 'react'
import { RiskDashboard, PortfolioChart } from '@/components'

export default function App() {
 const [portfolio, setPortfolio] = useState([])
 const [riskScore, setRiskScore] = useState(0)

 useEffect(() => {
  fetchPortfolio().then(setPortfolio)
  computeRisk(portfolio).then(setRiskScore)
 }, [portfolio])

 return (
  <RiskDashboard score={riskScore}>
   <PortfolioChart data={portfolio}/>
  </RiskDashboard>
 )
}
9:41
foliotrack
Total Portfolio
$42,810
▲ +$1,204 (2.89%)
today
7-DAY PERFORMANCE
Allocations
BTC
38%
Equities
31%
Gold
18%
Cash
13%
Home
Markets
Alerts
Profile
9:41
● LIVE
RiskView
HIGH
COMPOSITE RISK SCORE
6.2 /10
MODERATE-HIGH
Daily P&L
VaR (99%)
-$2,140
Sharpe
1.42
Max DD
-12.3%
Beta
0.84
BTC concentration above threshold
Dashboard
Risk
Models
Reports
9:41
✓ HMRC
taxwise
2024–25
ESTIMATED TAX LIABILITY
£ 8,420
▼ £640 saved via planning
Income Breakdown
Personal Allowance
0%
£12,570 / £12,570
Basic Rate
20%
£21,200 / £37,700
Higher Rate
40%
£0 / £125,140
Deadlines
Self Assessment
31 Jan
VAT Return Q4
7 Feb
Overview
Tax
HMRC
Reports

Building models that
actually trade.

Random Forest Classifier
Live Testing
Ensemble model trained on 50+ technical and macro features — RSI divergence, volume profile, order flow imbalance, VIX regime, Fed funds rate delta. Outputs directional probability for intraday and swing trade setups with feature importance rankings.
73%
Test Accuracy
50+
Features
1.42x
Sharpe Ratio
LSTM Price Forecasting
Research
Long Short-Term Memory network trained on multi-timeframe OHLCV data with attention mechanisms. Captures temporal dependencies across 60-bar lookback windows. Tested on BTC/USD, ETH/USD and SPY with walk-forward validation to prevent data leakage.
3.8%
MAPE
60bars
Lookback
3assets
Tested On
Momentum + Mean Reversion
Backtested
Hybrid strategy combining ML-filtered momentum signals with Ornstein-Uhlenbeck mean reversion on co-integrated pairs. Regime detection via Hidden Markov Models to switch between strategies depending on VIX and correlation environment.
2.1x
Sharpe (BT)
4yr
Backtest Window
HMM
Regime Model
Custom Indicator Library
Python
Proprietary Python library of 20+ indicators beyond standard TA-Lib: adaptive ATR bands, multi-timeframe confluence scores, liquidation cluster proximity index, and on-chain metric integration (NVT, SOPR) for crypto-specific signals.
20+
Custom Indicators
On-chain
Data Sources
1
Data Ingestion
OHLCV via CCXT · On-chain via Glassnode API · Macro via FRED · Sentiment via Reddit NLP
CCXT Glassnode FRED API
2
Feature Engineering
50+ features: RSI divergence, volume delta, ATR bands, order flow imbalance, VIX term structure, liquidation proximity
TA-Lib pandas NumPy
3
Model Training & Validation
Walk-forward cross-validation with purged K-fold to prevent leakage. Hyperparameter tuning via Optuna Bayesian search
XGBoost Optuna scikit-learn
4
Backtesting
Position sizing via Kelly Criterion, slippage & commission modelled. Benchmark: Buy & Hold, Random entry. 4-year OOS window
Backtrader vectorbt
5
Live Signal Deployment
REST API serving signals to React Native app dashboard. Alerts via Telegram bot on regime shifts and high-confidence setups
FastAPI Telegram React Native
LSTM BUY 73% INPUT HIDDEN LSTM OUTPUT RSI VOL OBV VIX
— Strategy — Buy & Hold
+180% +90% +0% +212% 2021 2022 2023 2024 Sharpe: 2.1x Max DD: -14.2% Win Rate: 61.4% Trades: 847
Volume Profile
0.88
VIX Regime
0.74
RSI Divergence
0.68
Order Flow Imb.
0.61
Fed Funds Delta
0.52
ATR Bands
0.44
NVT On-Chain
0.36
Pred LONG
Pred SHORT
Act LONG
412
TP
118
FN
Act SHORT
94
FP
376
TN
Precision: 81.4% Recall: 77.7%
BULL BEAR BULL DRIFT Q1'21 Q1'22 Q1'23
Bull (52%)
Bear (21%)
Drift (27%)
scikit-learn TensorFlow Keras pandas NumPy TA-Lib Backtrader hmmlearn XGBoost SHAP vectorbt