Skip to content
Bilal Khan
  1. 01About
  2. 02Work
  3. 03Experience
  4. 04Photography
  5. 05Contact
Résumé ↗bilalkhaninbox2004@gmail.com
Arkyv — The archive of

Bilal Khan.

Software engineer working across quantitative research, machine learning, and systems.

Selected workRésuméGet in touch
Latest
Software Engineer, Helwa AI
Studying
Computer Engineering, Cal Poly SLO
Focus
Quant research · ML · Systems
Graduating
May 2027
01 — About

About

I'm most interested in the part of a system that tells you whether it works.

A backtest, a classifier, a processor — building the thing is usually the easy half. The hard half is knowing whether the result is real. So I build that half too: validation layers that make failure visible, research tools that won't let me quietly clean up my own record, and hardware I understand down to the multiplexer.

Away from the desk
02 — Work

Selected work

Research tools, models, and the occasional product. Three in depth, the rest in the index.

01 · Personal · Quant · 2026Active

falsify

An event-study harness that enforces pre-registration, so I cannot retroactively clean my own research record.

Why
Cheap backtests make it easy to find green equity curves in pure noise. falsify makes every attempt count.
My role
Designed the method and built the harness, trial log, and standalone viewer.
Result
Every filter that looked strong in exploration has so far flipped sign in confirmation — the discipline is the product.
Stack
  • Python
  • polars
  • NumPy
  • SciPy
  • DuckDB-WASM
  • Lightweight Charts
Read the case study
02 · Personal · Quant · 2026

Signal Stack

An indicator stack where every module answers exactly one orthogonal question, then gets AND-gated.

Why
Composite scores hide which condition failed. Separate, single-purpose modules make every decision explainable.
My role
Designed the five-module architecture and calibrated each threshold against empirical null distributions.
Result
A non-repainting stack: a signal that appears intrabar cannot quietly disappear before the close.
Stack
  • Pine Script v6
  • Python
  • NumPy
  • SciPy
  • TradingView
  • Dukascopy tick data
Read the case study
03 · Machine Learning · 2021

Parkinson's Detection from Hand Drawings

Classifying Parkinson's from spiral and wave drawings — four architectures, tried in sequence until one held.

Why
Tremor shows up in handwriting before it is easy to diagnose. A drawing is a cheap, non-invasive screening signal.
My role
Built and compared all four approaches, from hand-crafted HOG features to a dilated CNN.
Result
Settled on a regularized dilated-convolution network, evaluated on confusion matrices rather than accuracy alone.
Stack
  • Python
  • TensorFlow
  • Keras
  • scikit-learn
  • XGBoost
  • scikit-image
  • OpenCV
Read the case study

Index

8 more entries

03 — Experience

Where I've worked, and what I owned there.

    • Ran the build-versus-buy evaluation for the platform and wrote the analysis the decision was made on. We built.
    • Trade logging with domain-specific fields, so a result can be interrogated after the fact rather than taken on trust.
    • A signal validator that checks strategy output against its own stated assumptions.
    • A regime classifier, so performance reads per market condition instead of as one blended number.
    • A walk-forward runner — a strategy only tested on the window it was fitted to has not been tested.

    Why it is built this way

    A backtesting platform makes it cheap to try configurations. That is the risk, not the feature — enough variants selected greedily will produce a clean equity curve on noise alone. The validation layer exists to make that failure mode visible rather than to make results look good.

    Stack
    • Python
    • FastAPI
    • PostgreSQL
    • Next.js
    • Time-series
  2. Stack
    • Model Context Protocol
  3. Summer 2024RemoteBay Valley TechSoftware Engineering Intern[One line: what you built or owned.]
  4. Oct 2022 — Aug 2023RemoteNeuroleapDevOps Intern[One line: what you built or owned.]
04 — Toolkit

Toolkit

Grouped by what I use them for. The numbers point to the projects where each one did real work.

Languages

Quant & data

Machine learning

Systems & hardware

  • Processes & shared memory06
  • Lock-free synchronization06
  • RISC-V (RV32I)07
  • FPGA · Vivado07
  • Linux / WSL06

Product & services

05 — Photography

Off the clock.

[One line about your photography — what you shoot, where, and with what.]

  • 01 Plate 01

  • 02 Plate 02

  • 03 Plate 03

  • 04 Plate 04

  • 05 Plate 05

  • 06 Plate 06

  • 07 Plate 07

  • 08 Plate 08

  • 09 Plate 09

  • 10 Plate 10

  • 11 Plate 11

  • 12 Plate 12

Outside engineering

  1. 01

    Photography

    [One or two sentences on what you shoot and why.]

  2. 02

    [Interest]

    [One or two sentences — something that says who you are away from a keyboard.]

  3. 03

    [Interest]

    [One or two sentences.]