01
Software & integration
Custom applications, internal tools, APIs and systems integration.
- Custom applications
- Internal tools
- APIs and systems integration
Tensor Consulting 張量諮詢
Mexico · Macau
Software, data and AI,with scientific rigor.
We turn manual processes and scattered data into tools that help you run your business and make clear decisions.
We learn how your business works, define what is worth solving and build a solution your team can use. From diagnosis to software, with verifiable results.
Our team at a glance
M.Sc. × 3
Three master’s degrees in physics from UNAM. Quantitative training applied to your business.
Macau
Graduate studies in data science and fintech at the University of Macau.
24 h
Coordination across Mexico and Macau. Coverage and response times agreed for each project.
Areas of practice
Tools shaped by the problem.
Technology for real-world problems.
02
Data & automation
Data preparation and validation, dashboards and reports, automated workflows.
- Data preparation and validation
- Dashboards and reports
- Automated workflows
03
AI & predictive models
Risk and classification models, applied generative AI, document processing.
- Risk and classification models
- Applied generative AI
- Document processing
04
Quantitative methods
Statistics, simulation and numerical methods, financial modeling, decision support.
- Statistics and simulation
- Numerical methods
- Financial modeling and decision support
Software
Software built around one process.
We design and build applications for concrete operations, with business rules on the server, permissions and traceability. LendFlow, for lending businesses, is the first of our solutions.
Solution 01 · Lending businesses
LendFlow
Built by Tensor Consulting
Every loan clear.
Every payment recorded.
Your portfolio, due dates and payments in one workflow. LendFlow connects client records with lending, collections and business reporting.
- Organize collections. See installments due today, upcoming and overdue.
- Track every payment. Partial payments, numbered receipts and updated balances.
- Manage access. Permissions for administrators, collectors and accountants; voided payments retain their history.
- Understand your portfolio. Principal recovered, interest collected and transactions exportable to CSV.
The proposal is tailored to your operating rules. Scope and rollout are agreed with you.
A loan, broken down
- Initial principal
- $10,000
- Periods
- 06
- Rate per period
- 2%
| # | Payment | Principal | Interest | Balance |
|---|---|---|---|---|
| 1 | 1,785.26 | 1,585.26 | 200.00 | 8,414.74 |
| 2 | 1,785.26 | 1,616.97 | 168.29 | 6,797.77 |
| 3 | 1,785.26 | 1,649.30 | 135.96 | 5,148.47 |
| 4 | 1,785.26 | 1,682.29 | 102.97 | 3,466.18 |
| 5 | 1,785.26 | 1,715.94 | 69.32 | 1,750.24 |
| 6 | 1,785.24 | 1,750.24 | 35.00 | 0.00 |
- Total interest
- $711.54
- Total payments
- $10,711.54
From your workflow to a concrete proposal
- 01
Your lending rules
We review how you calculate interest, schedule due dates and record payments.
- 02
Examples that reconcile
We check installments, balances and receipts against agreed examples before rollout.
- 03
A written scope
We define configuration, initial data, training and support in the proposal.
Consulting
Decide with evidence
before you invest.
Not every problem needs a new system or an AI model. Consulting tells you what to do, in what order and at what risk — or checks whether what you already have, or what you’re being offered, actually works.
01
Diagnosis & roadmap
For companies that want to put their data, automation or AI to work and don’t know where to start — or that have several initiatives and need to prioritize them.
Deliverable. A diagnosis report and roadmap with prioritized initiatives: scope, required data, estimated effort and success criterion for each one. Includes a results presentation.
Explore the scope: Diagnosis & roadmap
What we do
- Interviews with the people who run the process and review of a data sample.
- Map of the current process: sources, manual steps, bottlenecks and risks.
- Improvement opportunities, including cases where AI doesn’t add enough.
- Prioritization by impact, effort, risk and data availability.
02
Technical review & second opinion
To validate a model, an analysis, a codebase or a vendor’s proposal before you invest in it or put it into production.
Deliverable. A findings report ranked by severity, with evidence and actionable recommendations. Includes a review session with your team or your vendor.
Explore the scope: Technical review & second opinion
What we review
- Methodology and assumptions: whether the approach answers the business question.
- Data: quality, bias and information leakage that inflates results.
- Figures and metrics: whether results reconcile and the chosen metric is the right one.
- Code and architecture: maintainability, traceability and operating cost.
Implementation
One solution for one
well-defined process.
Together we choose a process and agree on what the solution must do, what data it needs and how delivery will be verified.
What we deliver
- 01
Process discovery
Process map, review of sources and criteria to evaluate the improvement. If we already ran the diagnosis with you, we start from it.
- 02
Software & integration
An application or automation, with agreed business rules and access controls.
- 03
Data & KPI dashboard
Reproducible preparation, with explicit formulas, periods and sources.
- 04
Quantitative analysis
Findings, inconsistencies and recommendations backed by the data.
- 05
Documentation & training
Agreed code and documentation, a user guide and team training.
What we need from your team
- One person who owns the process and validates decisions and results.
- Access to a representative data sample and the agreed sources.
- Time to review progress and test the solution with real cases.
Our approach
Scientific rigor, in practice.
Physics training applied to business processes.
We met in the graduate physics program at UNAM. There we learned to frame open problems, model them and validate every result before accepting it. Today we apply that method to our clients’ processes.
Measure before we promise
Every recommendation starts from the data and a baseline.
Separate signal from noise
We tell a data error apart from a real pattern.
Keep everything verifiable
Findings come with evidence and steps that can be reproduced.
Say when it isn’t worth it
If a simpler solution is enough, we recommend it.
The team
Physicists who build
with data.
We bring together physics, business analysis and software engineering. The people who study your process also build and validate the solution.
- Mexico City
- --:--
- Macau
- --:--
Between Mexico and Macau
With 14 hours between our cities, we organize 24-hour operating coverage. Channels and response times are agreed for each project.
Mexico
Alejandro
Alejandro
Alvarado Corzo
Data & Software Engineering
A physicist by training (MSc, UNAM; PhD in progress), Alejandro integrates and validates financial and operational data and builds the software around it — from APIs to business rules — so every report adds up.
Experience & projects: Alejandro Alvarado Corzo
- Reconciles financial and operational data across databases, APIs, Excel and dashboards — so the numbers in every report add up.
- Turns end-to-end discrepancies into automated rules and validations that catch errors before they reach a decision.
- Full-stack engineering: frontend, backend, APIs and business rules.
- Co-author of an international scientific publication (2026).
Macau
Oscar
Oscar
Carranza Jiménez
Data Science (Financial Technology) & Computational Modeling
Oscar develops computational models and reproducible Python tools for simulation and data analysis. He holds an MSc in Physics from UNAM and is pursuing an MSc in Data Science (Financial Technology) at the University of Macau.
Experience & projects: Oscar Carranza Jiménez
- Numerical simulation and analysis of complex systems using Python.
- Independent credit-risk classification project using logistic regression and random forests.
- Engineering Physics graduate from UAM and recipient of its University Merit Medal.
Mexico
Axel
Axel
Morales Buendia
Data Science (Machine Learning) & Risk Analytics
A physicist by training (MSc, Institute of Nuclear Sciences, UNAM), Axel analyzes data and builds machine-learning models — such as XGBoost and random forests — designed for business impact, from ranking credit risk to reducing fraud losses. Results reach decision-makers through SQL and Power BI reporting.
Experience & projects: Axel Morales Buendia
- Fraud detection portfolio project on 6.3 million transactions: caught 80% of fraud losses while flagging only 0.28% of transactions.
- Credit-risk portfolio project on 150,000 borrowers: an XGBoost model (ROC-AUC 0.87) whose riskiest 10% captures 5.5× more defaulters than random review.
- SQL and Power BI: turns raw data into executive dashboards and KPIs that drive decisions.
- Computational physics: cut a numerical relativity simulation from more than three months to five days, and spent two years on large-scale data at the HAWC observatory.
How to engage
From first call to production.
Conversation → Consulting → Implementation → Support
- 01
Initial conversation
We learn the process, the problem and the data available — and tell you whether we can help.
- 02
Consulting
Diagnosis or technical review, with its own scope and budget.
- 03
Implementation
Quoted by deliverable, with written acceptance criteria.
- 04
Support & evolution
Agreed separately, based on how the solution is used.
Whatever the size of your business
Our initial scope is deliberately small — one process, up to two data sources, up to five KPIs — so you can start with a focused budget and grow from there. Startups, family businesses and established firms get the same rigor.
If the process is already defined, you can go straight to implementation. Price and timeline are set after reviewing the process and a data sample.
We treat your information as confidential and can sign a non-disclosure agreement before receiving any data.
Contact
Let’s start with the process
you want to improve.
Tell us what you need to decide or fix. We reply with next steps — or tell you honestly if we’re not the right fit.
To get started, tell us:
- Which process you want to improve.
- How you handle it today.
- What outcome you need.
Opens your email app with a short guide to your first message.