Research-driven data science

Complex data. Defensible answers.

Quantum Data Science brings scientific reasoning and engineering rigor to complex data problems—from biomedical research and statistical modeling to AI systems, process automation and operational decision support.

We design studies, build analytical pipelines and develop validated models and data products for organizations that need more than a black-box answer.

Conceptual illustration: regulatory networks, expression matrices, statistical curves and data pipelines converging into models and systems

Research

Research & Life Sciences

Bioinformatics, omics, experimental design, statistical inference and computational support for research groups, laboratories and healthcare organizations.

Leukocyte fraction across TCGA tumor cohorts, with adrenocortical carcinoma highlighted
Original research · Adapted from Muzzi et al., Front. Endocrinol., 2021 · CC BY

Organizations

Organizations & Operations

Predictive models, process automation, document intelligence, system integrations and data products for operational and strategic decision-making.

3D cost surface over two decision variables with the optimum highlighted
Applied project · Anonymized optimization analysis developed for a client.

The analytical visualizations presented throughout this page originate from research, academic and applied projects developed by João Muzzi. Client-related materials have been anonymized to preserve confidentiality.

The question comes before the tool.

Every project starts with a clearly defined question or operational objective. We then select the methods, models and technology appropriate to the problem—not the other way around. The result is work that can be reproduced, examined and used with confidence.

Mean ocean-current velocity field around an archipelago
Academic project · Ocean-current velocity field around the Philippine archipelago, estimated from a public flow dataset. Developed in the MITx Statistics and Data Science MicroMasters program.

Services

What we do

Five lines of work, combined as each project requires.

01

Bioinformatics & omics analysis

RNA-seq, microarrays, transcriptomics, regulatory networks, gene-set enrichment and variant interpretation. End-to-end, reproducible pipelines from raw data to interpretable results.

02

Statistical modeling & experimental design

Study design, sample-size and power, survival analysis, mixed models, Bayesian methods. Correct inference, honest uncertainty, results that hold up under peer review.

03

Machine learning & AI systems

Predictive and risk models, classification, language-model workflows for documents and reports. Validated the way science demands: proper splits, calibration, and a clear account of what the model can and cannot do.

04

Data products & process automation

Dashboards, APIs, automated reports, system integrations and data pipelines that connect to the systems you already use, from spreadsheets to ERPs and laboratory systems.

05

Scientific & research support

Analysis for manuscripts and grant proposals, publication-quality figures, methods sections, reproducibility audits and training for research teams.

Not sure where your problem fits?

Most projects touch more than one line. Send us a description of your data and your question, and we will tell you what a sound approach looks like, and whether we are the right partner for it.

Describe your project →

Approach

How we work

Scientific rigor

Explicit hypotheses, appropriate methods, and uncertainty reported rather than hidden. We say what the data supports and where it stops.

Reproducibility

Versioned code, documented pipelines and environments that run again next year. Everything we build is handed over; you own it.

Engineering discipline

Models and pipelines built to run in production: tested, monitored, integrated with the systems your team already uses.

Clear communication

Plain-language summaries for decision makers, full technical appendices for those who want to check every step.

How a project runs

  1. DiscoveryA conversation about your data, your question and what a useful answer would change.
  2. ProposalScope, methods, deliverables, timeline and cost in writing, with no ambiguity.
  3. ExecutionShort cycles with checkpoints, so direction can be adjusted early rather than late.
  4. HandoverCode, documentation and a walkthrough. Support afterwards if you want it.

Track record

Selected experience

Fields in which our team has delivered research and applied work.

  • Computational oncology, bioinformatics and multi-omics research
  • Statistical modeling and experimental design
  • Predictive modeling and AI applications in healthcare
  • Data automation for legal, financial and operational processes
  • Reproducible analytical pipelines and production data systems
  • Peer-reviewed research and international collaborations
Regulon reconstruction in adrenocortical carcinoma: hub regulator with positive and negative targets, and their positions along chromosomes
Original research · Adapted from Muzzi et al., Cancers, 2022 · CC BY
João Muzzi

Founder

João Muzzi, PhD

Founder & Computational Researcher

Quantum Data Science was founded by João Muzzi, PhD in Microbiology, Parasitology and Pathology from the Federal University of Paraná, holder of the MIT MicroMasters in Statistics and Data Science and a bachelor's degree in Bioprocess Engineering and Biotechnology.

He is a postdoctoral researcher in computational oncology, author of thirteen peer-reviewed publications with international collaborations, and an elected full member of Sigma Xi, the scientific research honor society.

His academic track record and earlier professional work are documented on his personal page. Quantum Data Science is the company through which that expertise is now offered to research groups and organizations.

Full profile at jmuzzi.quantumds.tech →

Contact

Bring us the question behind your data.

Tell us what data you have, what you need to determine and how the answer will be used. That is enough for an initial assessment. We reply within two business days.