ML engagements

Submit files.Scope milestones.Pay on approval.

Upload files. Get a fixed quote. Fund milestones in escrow. Pay only when you approve the work.

Files in
Fixed price
Escrow
Review first

Know your data first.

Fixed

Price upfront

No hourly billing — you know the full cost before work starts.

Sample output

Audit → clean → report

Pay on approval

Fund milestones upfront

Release when you approve

Get a quote

Every milestone ships something tangible

Handover-ready outputs at every gate — not slide decks or hours logged.

Written report

PDF summary stakeholders can read without opening code.

Reproducible notebook

Documented analysis you can rerun, audit, and extend.

Charts & visualizations

Plots, embeddings, and metric dashboards tied to your data.

Handover memo

What we did, what it means, and recommended next steps.

EDA

Sample output

Missing values · dtypes · distributions

Unsupervised

Sample output

UMAP · cluster labels · segment sizes

Supervised

Sample output

AUC 0.91ROC curve · precision/recall · SHAP

How it works

Fixed price, escrow-backed, review at every gate — from files to finished work.

01

Submit files & brief

Upload your datasets and describe the problem. We review scope and data quality.

02

Receive a quote

Fixed-price milestones with line-item costs and deliverables. Accept the full engagement, one phase, or decline — no obligation.

03

Fund milestones

Work is split into milestones. You fund each one before we begin that phase.

04

Review & release

Internal QA, then your approval. Payment is released only when you sign off.

What to expect

Review first, fixed quote next, work begins when you fund.

Day 1-2

Initial review

We review your submission, assess data quality, and prepare scope questions.

Day 3-4

Quote delivery

You receive a fixed-price proposal with milestones, costs, and deliverables — accept, decline, or fund one phase.

Day 5+

Work begins

Once you accept and fund the first milestone, we start the engagement.

The review flow

How work moves from analyst to you.

Analyst completes work

Internal QA review

Revisions if needed

Your approval

Payment released

Our guarantees

Commitments we make to every client.

1

Quality guarantee

If a deliverable does not meet the agreed specifications, we rework it at no additional cost.

2

Refund policy

If we cannot deliver the agreed scope, unused milestone funds are refunded in full.

3

Data security

All data is encrypted in transit and at rest. We delete client data within 30 days of project completion unless otherwise agreed.

Your money stays in escrow until you approve.

Fund milestones upfront. Internal QA runs first — then you review polished deliverables. Unused funds are refunded if we cannot deliver the agreed scope.

Start a project

Milestone escrow

Pay per phase — never for unfinished work. Release payment only when you sign off.

Two-gate review

Analyst delivery → internal QA → your approval. You only see work that passed both gates.

Sample engagement

Illustrative only — every project is scoped and quoted individually.

Retail
Unsupervised
3 weeks

Customer segmentation engagement

Cluster profiles, embedding visualizations, and a recommendations memo — scoped as three fixed milestones.

Feature engineering & prep

Clustering exploration

Visualization & recommendations

From KES 250,000

Starting points in KES — final quote depends on data size, complexity, and scope.

Get your fixed-price quote
UMAP · cluster labels · segment sizes

Services

EDA, unsupervised, and supervised learning — fixed-price, scoped as milestones.

Exploratory Data Analysis

We profile, clean, and summarize your datasets — surfacing patterns, quality issues, and actionable insights in a clear report you can share with stakeholders.

Typical milestones

1

Data audit & schema review

2

Cleaning & validation

3

Profiling report

Missing values · dtypes · distributions

Sample output

Get a fixed quote

Deliverables

  • EDA notebook
  • Summary report (PDF)
  • Cleaned dataset

Use cases

  • Pre-modeling data quality assessment
  • Executive data summaries for stakeholders
  • Identifying data gaps before collection
  • Audit trail for regulatory compliance

Tools

pandas
ydata-profiling
matplotlib
seaborn
Great Expectations

Compare service types

Pick the engagement that matches your goal.

EDA

Data cleaning & validation

Statistical profiling

Feature engineering

Clustering & segmentation

Anomaly detection

Model training & tuning

Evaluation metrics & reports

Production-ready artifacts

Unsupervised

Data cleaning & validation

Statistical profiling

Feature engineering

Clustering & segmentation

Anomaly detection

Model training & tuning

Evaluation metrics & reports

Production-ready artifacts

Supervised

Data cleaning & validation

Statistical profiling

Feature engineering

Clustering & segmentation

Anomaly detection

Model training & tuning

Evaluation metrics & reports

Production-ready artifacts

EDAUnsupervisedSupervised
Data cleaning & validation
Statistical profiling

Feature engineering

Clustering & segmentation

Anomaly detection

Model training & tuning

Evaluation metrics & reports
Production-ready artifacts

Transparent, milestone-based pricing

No hourly billing. Starting points in KES — final quote depends on data size, complexity, and scope.

Fixed

Price upfront

0

Hourly billing

2-gate

Quality review

100%

Escrow-backed

Sample EDA engagement

Typical scope for a mid-size dataset — quoted individually after we review your files.

M1

Data audit & schema review

M2

Cleaning & validation

M3

Profiling report & handover

From KES 150,000

Starting points in KES — final quote depends on data size, complexity, and scope.

Get your fixed-price quote

What's included

  • Fixed-price milestones with no hourly billing

  • Internal QA review before client delivery

  • Source code and notebooks (yours to keep)

  • Documentation and handover sessions

  • Two rounds of revisions per milestone

  • 30-day post-delivery support window

Payment methods

M-Pesa
Bank transfer
Card via Paystack
Wire transfer

About Takwimu Labs

Takwimu means statistics in Swahili. We run ML engagements the way they should work — scoped, reviewed, and paid fairly.

Our mission

ML engagements done properly.

Too many data-science projects drift on scope, skip review, or bill opaque hours. Takwimu Labs fixes that with a structured flow: you submit files, we quote, you fund milestones, we deliver, you approve, payment releases.

We do not host ML pipelines or sell compute. We exchange artifacts — datasets in, models and reports out — with clarity at every step. Our goal is to make machine learning accessible, predictable, and fair.

Transparency

No black boxes. We explain our methods, share our code, and price work clearly.

Quality over speed

We would rather deliver late and correct than fast and flawed. Internal QA ensures standards.

Client ownership

Your data, your models, your code. We build to hand over, not to lock in.

Fair pricing

Fixed milestones mean you pay for outcomes, not hours. We share the risk.

Who does the work

Specialist teams behind every engagement.

AT

Analytics Team

Data Scientists & ML Engineers

Our analysts have shipped ML in production across fintech, retail, and healthcare. They handle EDA, feature engineering, and model development.

Q&

QA & Review

Senior Reviewers

Every deliverable passes through internal QA before reaching you. Reviewers check methodology, code quality, and documentation.

CS

Client Success

Project Coordination

Your point of contact throughout the engagement. They manage timelines, handle communication, and ensure smooth handovers.

Why work with us

No hourly billing, no scope drift — just fixed milestones and review-gated payments.

Milestone-based pricing

No hourly billing. Each phase is scoped and priced upfront — you know the cost before we start.

Two-gate quality review

Internal QA catches issues before you see them. You only review polished deliverables.

Escrow-backed payments

Funds are held securely. We only get paid when you approve the work.

Clear deliverables

Notebooks, reports, models — everything documented and ready for your team to use.

No lock-in

We hand over artifacts you own. No proprietary platforms, no ongoing dependencies.

Domain expertise

Our analysts have shipped ML in finance, retail, healthcare, and logistics.

Who this is for

Domain-aware ML engagements — and what to send when you reach out.

Industries we serve

Financial Services

Credit risk scoring, fraud detection, portfolio analytics

E-commerce & Retail

Customer segmentation, demand forecasting, basket analysis

Healthcare

Patient stratification, operational forecasting, outcome analysis

Logistics & Supply Chain

Route optimization signals, demand planning, anomaly detection

Telecommunications

Churn prediction, network usage patterns, customer lifetime value

Agriculture & AgriTech

Yield forecasting, sensor anomaly detection, supply forecasting

Great fit

Teams with datasets ready to analyse or model

Leaders who want fixed-price milestones, not hourly billing

Organisations that need documented, handover-ready deliverables

Not the right fit

· Production MLOps, real-time serving, or full platform builds

· Projects without a clear business question or success metric

· Engagements requiring 24/7 on-call or embedded staff augmentation

What to prepare

Dataset files (CSV, Parquet, JSON) or secure cloud access instructions

A brief: business problem, desired outcome, and timeline

Column or data dictionary if available

Known data quality issues, constraints, or compliance requirements

NDA signed before any sensitive data is transferred

Common questions

Everything you need to know before starting a project.

Still have questions?

EDA projects typically take 1-2 weeks. Unsupervised and supervised learning projects range from 2-6 weeks depending on complexity and data volume.

We work with CSV, Parquet, JSON, SQL exports, and most common tabular formats. For larger datasets, we can set up secure transfer via cloud storage.

Yes. We sign NDAs before any data is shared. Your data privacy is paramount — we never use client data for training our own models or share it externally.

Minor revisions within the original scope are included. Significant changes or new requirements are scoped as additional milestones with separate quotes.

Absolutely. We often collaborate with in-house data teams, providing specialist capacity for specific projects while your team handles production deployment.

We assess data complexity, volume, and project scope. Larger datasets or complex feature engineering increase the cost. You get a fixed quote before any work begins.

No. The milestone price is the full price. We do not charge for compute, storage, or platform access. You pay for scoped work, nothing else.

M-Pesa, bank transfer, and card payments via Paystack. Quotes are issued in Kenyan shillings (KES).

The milestone model is already pay-as-you-go. You fund each milestone before that phase begins, so you never pay for unfinished work.

If requirements change mid-project, we pause and re-quote. You approve the new scope and cost before we continue — no surprise bills.

We discuss the issues, make revisions, and resubmit. If we cannot reach agreement, unused funds are refunded and the engagement ends cleanly.

Yes. If you need to pause between milestones, we can hold your slot for up to 30 days. Longer pauses may require re-scoping.

You do. All code, models, and reports are yours to use, modify, and deploy without restriction.

We sign NDAs, use encrypted transfer, and can work within your cloud environment if required. We never use client data for internal training.

Ready for a fixed-price quote?

Share your dataset and goals — we respond with a scoped proposal, milestone breakdown, and fixed costs.

Start a project

No hourly billing · Escrow-backed · Usually within 48h

Start a project

Ready to work with your data?

Tell us about your dataset and goals. You'll get a fixed-price proposal with milestone breakdown — we typically respond within 48 hours.

Fixed-price quote

Line-item milestones with fixed costs. No hourly billing, no surprise invoices.

Two-gate quality review

Internal QA before you see it. You only approve polished deliverables.

Share your dataset

Upload files securely after we confirm scope and send onboarding details.

Dedicated support

Your point of contact throughout the engagement.

Industries we serve

FinTech
Retail
Healthcare
Logistics
AgriTech

Share your brief here first — file uploads follow after scope is confirmed. Fixed-price proposals typically arrive within 48 hours.

Data security

All data is encrypted in transit and at rest. We delete client data within 30 days of project completion unless otherwise agreed. We sign NDAs before sensitive data is shared and can work in your cloud environment if required.

Email us directly

hi@piviotech.com

Initial response

Usually within 48h

Based in

Nairobi, Kenya

ProcessServicesPricingAboutAudienceContact