Deep Wave Research

Independent research lab · Established 2019

Deep Wave Research

An independent research lab I founded and operate, working across signal processing, machine learning, and complex systems.

Established in 2019. I worked as a freelance Consultant from May 2019 to February 2024, and as Research Engineer from March 2024 to the present. The work now spans scientific software, quantitative finance, and climate forecasting.

The lab

An independent research practice.

Deep Wave Research is the independent lab I founded in 2019. Unless a collaborator is named on a project page, the research design, software, experiments, and technical reports are my independent work.

I choose projects that require both analysis and implementation. Depending on the question, that has meant building data pipelines, reproducing published methods, training forecasting models, writing GPU code, or designing controlled comparisons.

I publish public technical reports, including negative results. Most underlying research code is not publicly released; software pages state their release and verification status explicitly.

My usual sequence is simple: define the question, build the method, test it on real data, and publish what happened.

The arc

How the work has changed since 2019.

I began in markets and gradually moved into broader signal-processing and forecasting problems.

  1. May 2019 – February 2024

    Founding and freelance consulting

    I established Deep Wave Research in 2019 and worked as a freelance Consultant on exploratory research, software, forecasting, and signal-processing prototypes.

  2. March 2024 – Present

    Research engineering

    My Research Engineer role brings together question formulation, research software, experiment systems, statistical evaluation, and technical reporting.

  3. 2022 – 2026

    Market studies

    I published standalone studies on microstructure, event studies, risk signals, cross-sectional factors, and model evaluation.

    Selected Topics in Quant Finance

  4. 2025 – 2026

    Distributional forecasting

    I shifted from predicting market direction to forecasting return distributions, path excursions, and volatility.

    Regime-Adaptive Probabilistic Forecasting

  5. 2026

    Departure from markets

    I applied the wavelet and forecasting work to tropical-Pacific sea-surface temperature, moving the research beyond finance.

    Wavelet-Domain ENSO Forecasting

The work

Research programmes and software

Selected programmes sit alongside native forecasting software. The technical records state results, limits, and release status.

December 2023 verification beside the lead-4 deterministic and 32-member-mean forecasts from the complex-wavelet Earthformer ENSO study
December 2023 verification beside the lead-4 deterministic and 32-member-mean forecasts from the complex-wavelet Earthformer ENSO study

2026

Wavelet-Domain ENSO Forecasting

Eleven linked documents examine tropical-Pacific sea-surface-temperature forecasting in an exactly invertible complex-wavelet space. They comprise a standalone manuscript, a measurement charter, and nine reports running from system validation to a live-deployment handoff.

Methods
Dual-tree complex wavelets, Earthformer cuboid attention, CMIP6 forecast pretraining, sequence diffusion, EMOS and ECC-Q calibration, hash-gated sealed testing, moving-block bootstrap governance
Result
Threads the research programme together without collapsing its evidence tiers. The record includes one pristine sealed-test win, two controlled studies that narrowed the design space, a probabilistic stack whose validation advantage only partly transferred, and a deployment contract for a prospective live test.
Explore research series

Research software · Pre-release

Framework-Free C++/CUDA Chronos and Moirai-MoE

Native implementations of two time-series foundation-model families, including forward and backward kernels, checkpoint loading, and fine-tuning without a tensor-framework runtime.

Pre-release; source not public. Small-variant forward checks demonstrated; broader fidelity and performance tests remain open.

Methods
C++20, CUDA, checkpoint conversion, per-tensor numerical comparisons, explicit fidelity gates, native fine-tuning
Result
Forward agreement demonstrated for small model variants. Base-variant checks and performance benchmarks remain unrun; full numerical equivalence is not claimed.
Read software record at Deep Wave
Validation pinball skill for learned heads with and without causal tabular features, from the frozen-backbone head lab
Validation pinball skill for learned heads with and without causal tabular features, from the frozen-backbone head lab

2026

Regime-Adaptive Probabilistic Forecasting

A staged programme studies distributional forecasting of BTCUSDT perpetual futures. It begins with frozen foundation-model embeddings, maps where the series carries learnable signal in a pre-registered scan, and ends with a walk-forward comparison of from-scratch and foundation-model forecasters staged on those verdicts.

Methods
PyTorch, return and excursion quantile supervision, Chronos-T5 and Moirai-MoE backbones, LoRA fine-tuning, conformalized quantile regression, pre-registered verdict thresholds, purged walk-forward evaluation
Result
Establishes that the usable signal is distributional rather than directional and maps where it appears across timeframes and horizons. A small from-scratch encoder defines the quality-versus-cost frontier over the foundation-model lanes, while fine-tuning helps in proportion to the deficiency of each frozen representation.
Explore research series
Baseline scatter from the FOMC statement sentiment and 2-year yield reaction report
Baseline scatter from the FOMC statement sentiment and 2-year yield reaction report

2020 – 2026

Selected Topics in Quant Finance

Six standalone studies cover distinct questions in quantitative finance. They examine FOMC statement sentiment against a 2-year yield proxy, intraday spectral signatures, one-second order-book prediction, change-point detection in SPX realized volatility, cross-sectional intraday reversal, and a public-data replication audit of a real-time equity-ML paper.

Methods
Event studies, NLP sentiment scoring, signal processing, market microstructure, change-point detection, cross-sectional factor research, replication auditing, bootstrap and significance testing, Python
Result
A set of careful, reproducible studies, each an honest result about what does and does not hold in intraday, cross-sectional, and risk-signal data.
Browse research collection