Rabhya smiling over her shoulder at golden hour

hi, I'm Rabhya

The one who asks “wait, how does that work?” and then helps build it.

I started as an engineer at Morgan Stanley, fell in love with financial markets, and went to Imperial to learn them properly. Scroll for the story, the ideas I have played with, and the things I do when I am not working.

Available now · London
Financial Marketsprices, portfolios RiskVaR, ES, stress DataSQL, dashboards MachineLearning me
  1. 2018 – 2022BTech, Computer ScienceSRM Institute of Science and Technology
  2. 2021Data Analyst InternAntPod, an agritech startup
  3. 2022 – 2024Intern to Senior Software EngineerMorgan Stanley
  4. 2025 – 2026MSc Risk Management and Financial EngineeringImperial College London
  5. nowI am hereLooking for my next role
02 · the story

How it started, and where it went

Chapter 1 · age 14
14

The year I wrote my first line of code.

I always loved maths. At 14, I found out it could run on a computer, and I was hooked. I was often the only girl in the room, and I never thought twice about it.

// age 14, probably
public class Hello {
  public static void main(String[] a) {
    System.out.println("hello, world");
  }
}

ICSE Class 10 turned out well. I finished first in school, and top of the class in 4 out of 6 subjects.

Chapter 2 · school, 11th and 12th

Physics, chemistry, maths and computers.

From there, school got bigger. I took Java first, then picked up C++. Java taught me to model the world with objects. C++ taught me what the computer is actually doing underneath.

Chapter 3 · undergrad

The place I tried everything.

Then came SRM, my sandbox. I studied Computer Science and Engineering and tried every corner: web apps, data visualisation, machine learning, and plenty of hackathons.

#1
Rank 1 in class
CGPA 9.71 out of 10, merit scholar
150+
Co-Founder, GirlUp Heron
helped 150+ women-led startups raise funds
2021
Data Analyst Intern
AntPod, an agritech startup: forecasting commodity yield volatility
Ops
Social Club SRM
Head of Operations, 2019 to 2020. Plantation drives and teaching children
three projects I liked ↓

Loan default prediction

Will this borrower repay? A machine learning model that reads credit history to predict it.

Research with my professor

Dimensionality reduction on huge datasets: hundreds of messy columns down to the few that carry the signal.

Thesis: algorithmic trading

Trading strategies built on technical indicators, with LSTM and ARIMA forecasts. It taught me to ask "can I prove it isn't luck?"

Chapter 4 · industry

Inside the engine room of finance.

After SRM, I joined Morgan Stanley as an intern and grew into a Senior Software Engineer, working across New York, Hong Kong, India and Montreal. I built the pipelines and monitoring behind the numbers, and learned that a number is only as good as the system that carries it.

  1. Market data

    raw prices arrive

    45Mdaily transactions
  2. Validation

    checks and rules, where a mis-keyed price like 1012.5 gets caught

    35%fewer pricing discrepancies
  3. Pricing

    price, Greeks, P&L

  4. Trading and risk

    the desks that use it

    25+systems fed by my APIs
101.25101.30101.281012.5 Market dataraw prices arrive45Mdaily transactions Validationchecks and rules35%fewer pricing discrepancies Pricingprice, Greeks, P&L Trading and riskthe desks that use it25+systems fed by my APIs caught!
a simplified picture: a price travels to the desks that rely on it, and a mis-keyed one gets caught on the way
where the curiosity came from

Day to day I was building the pipes: prices in, Greeks out, P&L explained. But the questions that stayed with me were the ones behind the numbers. Why does a price move? What does risk actually measure? Who decides what a number is worth? Those questions pulled me from engineering towards finance.

Chapter 5 · Imperial

Then I went back to learn the finance properly.

An MSc in Risk Management and Financial Engineering at Imperial College London, completed with Merit. These are the modules where I did my best work.

80%
Systematic Trading Strategies with Machine Learning
76.4%
Risk Management and Valuation
70.15%
Applied Project, my dissertation
Dissertation · 70.15%

Geopolitical risk-conditioned expected shortfall

Does a news-based geopolitical risk index sharpen tail-risk forecasts? I tested it on 97.5% expected shortfall against GARCH-EVT, filtered historical simulation and other models, across equities, oil, gold and US Treasuries, 2000 to 2026.

everything I studied

Trading and machine learning

  • Systematic Trading Strategies with Machine Learning
  • Applied Trading Strategies

Risk and valuation

  • Risk Management and Valuation
  • Financial Engineering
  • Stochastic Calculus for Finance

Portfolios and investing

  • Asset Allocation and Investment Strategies
  • Investments and Portfolio Management
  • Private Equity and Venture Capital

Data and foundations

  • Empirical Finance
  • Financial Statistics
  • Foundations in RMFE and Ethics

Beyond the classroom: Algorithmic Trading Society, Venture Capital and Private Equity Society, Women in Business Society, Portfolio Investment Analyst.

Which brings me here: looking for my next role in risk, quant or investing.

03 · concept lab

Concepts I enjoy, explained by playing

Along the way I kept turning ideas into things I could play with. Drag the sliders and watch each one move. Every card comes from my work or studies, simplified so the idea is easy to see.

Derivatives · Black-Scholes and the Greeks

An option's price has a personality.

Delta is how much it moves with the stock. Gamma is how fast delta changes. Vega is how much it cares about fear. Drag the sliders and watch the curve bend.

-Price
-Delta
-Gamma
-Vega per vol point
call option, strike 100, 2% rate. dashed line = delta
Put to workAt Morgan Stanley, Greeks calculation and real-time P&L attribution across rates, credit and equities ran on the pricing infrastructure I helped build.
Risk · expected shortfall in a crisis

VaR says where the bad days start. Expected shortfall says how bad they get.

My dissertation studied expected shortfall when geopolitical risk is high. Drag the stress slider: the tail fattens faster than the volatility does.

-daily VaR, 99%
-daily expected shortfall, 97.5%
-shortfall / VaR 99%
simulated returns, illustrative only. not my thesis results
Put to workMarket Risk Engine: a VaR and CVaR engine for a £10M portfolio. 99% VaR validated over 1,311 days with a 1.6% breach rate (Basel III, Kupiec POF), then stress-tested on COVID-19, the 2022 rate hikes and Russia-Ukraine: worst case -23.6% against the model's VaR. Code on GitHub
Risk · pricing an option, step by step

Price an option by walking a tree.

Each step, the stock can go up or down. Start from the payoffs at the end and work backwards, averaging as you go. Add steps and the answer homes in on the Black-Scholes price.

-tree price
-Black-Scholes price
-gap
-up move per step
stock = 100, one year, 2% rate. simplified on purpose.
Systematic trading with machine learning algorithms

Can a model find an edge, and prove it isn't luck?

The classic trap is peeking at the future. I tested 13 machine learning models with walk-forward testing: train on the past, test on the next unseen block, slide forward, repeat. Only what works on data the model has never seen counts.

train on the pasttest on the unseen
simulated returns to show the method, not my results
Put to workMy financial ML metamodel for systematic futures trading: triple-barrier labelling, HMM and GMM regime detection, 100+ alpha features, 13 models benchmarked. Out of sample: 12.3% CAGR, 1.42 Sharpe, -2.87% max drawdown. Code on GitHub
Systems · data you can trust

Every number downstream is only as good as the pipe.

A hundred ticks come in. Some are bad. Validation checks catch most of them before they reach pricing and risk. Slide the bad-data rate and watch what leaks through.

-bad ticks arriving
-caught by validation
-leaked downstream
0.92assumed validation recall
simulated ticks, to show the idea
Put to workAt Morgan Stanley I worked on pipelines handling about 45M daily transactions across 25+ systems, and cut discrepancies by 35%.
Private equity · valuation

A company is worth the cash it will make, in today's money.

That is a discounted cash flow, or DCF. Grow the cash, raise the discount rate, and watch where the value hides. Spoiler: mostly in the far future.

-enterprise value (£m)
-share from terminal value
-years 1 to 5 (£m)
-value / year-1 cash
starts from £100m cash flow. simplified
Put to workAlphaLens, an equity research platform I built. It automates the financial modelling and runs a DCF and comparable-company valuation of Ferrari N.V. behind a Buy, Hold or Sell call. Code on GitHub
Private equity · LBO

Buy a company with borrowed money, then let it pay the debt.

In a leveraged buyout the business's own cash flow pays down the loan. Returns come from three places: the company grows, the exit multiple moves, and the debt shrinks.

-equity in (£m)
-equity out, year 5 (£m)
-money multiple
-IRR
starts from £100m EBITDA, 5-year hold, simplified cash sweep
Put to workSpire Healthcare take-private: £2.36 a share at 8.5x EV/EBITDA, three debt tranches, net debt falling from 4.6x to 2.2x EBITDA. Base case 18.3% IRR and 1.91x MOIC, with bear and bull cases and the investment committee memo I wrote.
Venture capital · thesis and memo

Most startups fail. A few change everything.

Venture returns follow a power law: a handful of winners pay for all the losers. That is why a venture investor writes a clear thesis and a sharp investment memo before backing anything.

fund 1
-fund multiple
-companies that lost money
-best company
-best company's share of returns

Anatomy of an investment memo

simulated outcomes for 25 investments. not real fund data
Put to workPolaris Imperial Partners, a $750M Series A/B fund thesis on defence autonomy and advanced nuclear. I served as GP for nuclear and energy: deal sourcing, regulatory diligence, benchmarking against five incumbent funds, 22 to 25 positions, 50% follow-on reserves, 1.5x to 5x net MOIC.
Portfolios · asset allocation

Two thousand portfolios walk into a frontier.

Every dot is a random mix of four assets. The top edge is the efficient frontier: the most return for each amount of risk. Pick your appetite.

-expected return
-bonds / equities / EM / commodities
-return per unit of risk
4illustrative assets
illustrative inputs
Put to workVector Global Macro, a simulated $100M multi-strategy fund across rates, credit and FX: +4.74%, Sharpe 4.71, 300bps ahead of the HFRX macro index, sized with inverse-volatility risk parity.
04 · how my stack grew

One skill on top of another

Looking back, each chapter added a layer. Blue is code, green is data, ML and maths, peach is finance, yellow is engineering craft.

Java and C++
Calculus
Combinatorics
Problem Solving

School

Age 14 to 17
Python, Pandas and NumPy
JavaScript, React and Node.js
Matplotlib, Seaborn, Plotly
Power BI and Tableau
Linear Algebra, Probability and Statistics
Machine Learning and Neural Networks

BTech

Computer Science Engineering
Java and SQL
Python and Ext JS
Real-Time APIs
Grafana and Prometheus
Pricing, Greeks, P&L

Morgan Stanley

Intern to Senior Software Engineer
R, Excel and Bloomberg
Scikit-learn and ML for Trading
AI and LLMs
Derivatives and Options Pricing
DCF and LBO
Portfolio Theory, VaR and ES

Imperial

MSc
05 · off the court

The other half of me

When I am not working, I am usually moving, cooking or planning a trip. I played U-18 basketball and district-level tennis, trained in Bharatanatyam for eleven years and earned a Master's in it. I also love chess, painting and travel, and I make food content as @hungryawns.

Basketball

made 0 of 0

Tennis

Over the net, one bounce, back again. District level, under 16.

Bharatanatyam

Eleven years, and a Master's.

Play me

A quick game of chess

You are white. I am going easy on you. Tap a piece, then a dot.

Your move.

Sketches and paintings

Where I slow down: pencil, acrylic and a lot of patience.

Pencil portrait sketch of a woman with her hands on her face and two butterflies
Acrylic painting of a misty pine forest
Pencil sketch of a woman resting her face on her hand
my food channel

Hungryawns

Hungryawns is my little food corner on Instagram. I love food and I love travelling, so I bring recipes home from everywhere I go, then cook them, film them and turn each one into a tiny story. Part hobby, part creative lab, all appetite. Come hungry: @hungryawns

Stack of baked almond and oat cookies
Penne with broccoli, courgette and herbs
A bowl of chicken curry garnished with coriander
A burger cut in half
Toasted muffin sandwiches with strawberries, blueberries and blackberries
Spiced chicken bites with coriander in a blue bowl
Hummus with spiced pita crisps
06 · contact

Let's talk.

Hello from London,

That is the story so far, and I am available now for roles across risk, quant and investing. I speak English, Hindi, some German and a little Japanese. If anything here made you curious, I would love to hear from you.

Rabhya