Investment research / Research systems

I research investments and build the systems to test them.

My training is fundamental — valuation, diligence, and a P&L read line by line. My method is systematic: write the thesis as a rule, test it across regimes, price the costs, and document what would prove it wrong. I am looking for a seat in investment research, portfolio analytics, or systematic research where the output is a recommendation someone can act on — and a process someone can audit.

8 Systematic strategies built, tested and documented end to end
26yrs Of market data behind the research library, 1999 to present
$50M Investment decision supported through diligence and LBO modeling
M.S. Finance, University of Alabama · value investing and quantitative finance

02 / How I work

Four rules I do not trade around.

Every strategy, model and business I look at gets the same treatment. The process is the product; the ideas are replaceable.

01

Underwrite, do not forecast

A thesis is a claim about cash flows and what I pay for them, not a prediction about the mood of a market. If I cannot state the unit economics, I do not have a position — I have an opinion.

02

Costs before conclusions

Fees, spreads, taxes and my own attention are part of the model. A strategy that only works before costs is a story about a strategy. The same is true of a business that only works before overhead.

03

Build the machine

One good trade is luck; a repeatable process is a business. When something works, the next job is to make it run without me — in code, with an audit trail, so it can be inspected and improved.

04

Write it down

Method, assumptions and failure modes get documented before capital moves. A process nobody can audit is a process nobody should trust — including me, six months from now.

03 / What I bring

Finance, business, engineering — and the writing that holds it together.

Full background
Financial

The numbers

Valuation, LBO and DCF models, diligence, and a P&L read line by line. I came up through value investing and private equity diligence, which is why I ask where the cash comes from before I ask what the chart does.

Business

The judgment

Sizing, risk budgets, drawdown tolerance, and the difference between a good business and a good price. Most of the return comes from what you refuse to own and how much you refuse to hold.

Engineer

The machine

Python, pandas, FastAPI, vectorbt, LEAN. I build the data layer, the backtest engine and the automation myself, so the research is reproducible by someone who is not me.

Writer

The record

Every idea ends in a document with its assumptions exposed: strategy write-ups in a fixed house format, company reports with a target and a disconfirmation list, build logs that include what failed.

04 / Now

What I am building this quarter.

Ideas are cheap until something ships. These are the workstreams currently moving, each with a public trail on this site.

Follow the work
  1. 01 Systematic research libraryEight strategies on 26 years of data, validated walk-forward, published with the failures included — including a meta-labelling model that did not work.
  2. 02 QuantConnect / LEAN portRebuilding the research stack as a LEAN algorithm with point-in-time universes and realistic fill and fee models. Logged publicly.
  3. 03 Equity research libraryCompany notes written to one standard: business quality, normalized earnings, embedded expectations, and what would prove the thesis wrong.
  4. 04 Open to the right seatInvestment research, portfolio analytics, or systematic research — roles where fundamental judgment, data and code meet.

“The edge is rarely the idea. It is the machine that executes the idea at scale, and the discipline to keep running it when it is uncomfortable.”

Working principle · 2026

If you are hiring for investment research, portfolio analytics, or systematic research, let’s talk.

I am open to full-time seats where fundamental judgment, data and code meet. Strategy write-ups, code and equity research are available on request.

sohaibmali@outlook.com