Valuation, Portfolio,
and Risk Models

01

MARKOWITZ PORTFOLIO
OPTIMIZATION

In-sample, a 14-stock portfolio reached a 1.66 Sharpe ratio vs 0.71 for the S&P 500

Built a mean-variance optimizer in Python using scipy's SLSQP solver, running 100,000 simulations to map the efficient frontier and solve Min Variance and Max Sharpe portfolios. Backtested both against SPY, QQQ, and XLK over 6 years.

02

FIXED INCOME VALUATION,
DURATION HEDGING & CONVEXITY

Duration hedge cut a $131,700 rate-shock swing to near zero, leaving an $8,560 convexity loss at ±3%

Modeled bond pricing under flat-yield and term-structure discounting, then built a duration hedge offsetting a $1M bond position with a short in a second bond. Added convexity adjustments to show where linear hedging breaks down.

03

MONTE CARLO CORPORATE EARNINGS RISK (C-VAR)

Cut tail earnings risk from 8.89% to 3.04% via targeted hedging

Simulated a firm's pre-tax earnings using 4,000 Monte Carlo scenarios across FX, commodity, and rate exposures, then measured downside using 95% C-VaR. Aluminum price hedging delivered the largest risk reduction.

04

DISCOUNTED CASH FLOW (DCF)
VALUATION DASHBOARD

Interactive dashboard that prices any stock and stress-tests it 20,000 ways

Built a live DCF model in Python that pulls real-time financials and auto-calculates WACC via CAPM. Generates Bull/Base/Bear scenarios and a 20,000-iteration Monte Carlo simulation of intrinsic value.

05

SELL-SIDE EQUITY RESEARCH: TERADYNE (NASDAQ: TER)

Buy rating at $89.79 target, later validated by 100%+ appreciation

Produced an institutional-style equity research report using Bloomberg Terminal data, valuing the stock via EV/EBITDA against peer comparables. Combined with DuPont analysis, Porter's Five Forces, and ESG scoring.

06

MACROECONOMIC DRIVERS
OF THE S&P 500

Log model explained 92% of S&P 500 variation across 251 months

Ran OLS regressions in levels and in logs on monthly macro data from 1990 to 2023 (GDP growth, CPI, unemployment, FDI). CPI and unemployment were the strongest statistically significant drivers in both versions.

07

CORPORATE CREDIT RISK
ASSESSMENT MODEL

Distinguished investment-grade from distressed credit across 10 companies in 10 sectors

Built a SQL + Python pipeline that scores 10 public companies across 11 leverage, coverage, liquidity, and cash flow ratios, rolling them into a weighted composite score and an AAA-to-CCC letter rating. Layered in sector-adjusted liquidity thresholds and covenant stress tests that cut EBITDA by up to 47.3%, the size of a real decline in the sample.

08

DIVIDEND PAYERS VS NON-PAYERS:
S&P 500 RETURN AND RISK BACKTEST

Non-payers returned 22.1% a year vs 15.0% for payers, 2018 to 2025

Split the S&P 500 into 401 dividend payers and 91 non-payers, built an equal-weight portfolio of each, and compared return, volatility, Sharpe ratio, and drawdown. Documented the survivorship and yield-snapshot biases as limits.

09

PORTFOLIO RISK MEASUREMENT
USING VAR AND CVAR

Backtested 3 VaR methods on 2024: 1 to 2 breaches at 95% vs 12.5 expected

Measured daily downside risk for a 30-stock portfolio with historical, parametric, and Monte Carlo VaR and CVaR, fitted on 2019 to 2023 and tested on 2024. All three methods overstated risk.