This site and the matching algorithm engineering workbench are operated by YI ALGORITHMS LLC, a company registered in the United States at 4277 W Watchmen Way, Herriman - 84096-2759, United States (US). The Company designs optimization models, scheduling and routing systems, data pipelines, forecast layers and simulations for operations teams. The developer behind the workbench is YiAlgorithms, and the name appears here because visitors should know exactly who builds and secures the software that may touch their information.
When this policy says the Company, YI ALGORITHMS LLC or we, it means the registered entity named above. When it says the workbench, the bench or the platform, it means the engineering and delivery environment this Company maintains for model building and analysis.
The engineering team that reads your problem note is small and deliberate. A handful of designers and engineers own every line that touches your data, which means responsibility is never scattered across a hundred handoffs. Each person who can open a model file or reply to a client thread is known by name and role, and each is bound by the same duties this page describes. That closeness is one reason the bench works: the same people who derive your schedule also guard the record that feeds it.
The Company operates from a single conceptual workspace called the bench rather than from many disconnected tools. Records arrive at known doors, move along known paths and end in known cabinets. Because the map of the flow is simple, the map of your rights over that flow is simple too: you can trace what was received, what it was used for and where it will stop being kept. Simplicity of architecture is a privacy feature in itself, and the Company keeps it on purpose.
This policy explains how the Company treats information in three separate roles. First, visitors to this website who browse, read or submit the contact form. Second, prospective clients who send a problem note or a scoping enquiry. Third, the operations teams of current clients whose aggregated or depersonalized operational data enters a modelling exercise on the bench.
The Company keeps these three streams separate in its record keeping. A browser visit should never be mixed with a client model. Where a visitor becomes a client later, the Company marks the boundary and applies the client terms to any project data from that point forward.
On the contact page a visitor may provide a name, an email address, a subject line and a free text problem note. The Company receives that exact content through the mail channel chosen at submission. Where a visitor calls the number shown, the Company may note the number and the reason for the call for follow up.
During client engagements the Company may receive operational records that a team consents to share for a stated derivation task. These may include volumes, timestamps, asset identifiers, zone codes or plan histories. The Company treats any item that could identify a person or narrow a specific shipment or shift as sensitive project material under the client agreement, not under this public policy alone.
The American legal framework uses many overlapping words for the same careful behaviour, and this policy tries to be plainer than the statute books. Personal data, personal information, personally identifiable information and protected data all mean roughly the same thing here: any record that can reasonably point to a specific living person. Operational data, by contrast, often describes trucks, zones, lines and shifts where no single person sits at the centre at all.
Where a model input is a route without a driver name or a roster without a worker identity, the Company still guards it as client property, but its privacy weight is far lighter than a record holding a salary or a health note. The heavier the record, the tighter the cabinet it goes in, and the stronger the reason the Company requires before it is used. That graduated care runs through every line of this page and every clause of the connected terms.
The bench follows a rule of least material: for a routing problem send the stops, the windows and the fleet; leave out names, home addresses and payroll notes unless a rule genuinely needs them. When the questions that matter can be answered with less identifying detail, the Company asks for less, because a record the Company never held is a record the Company can never lose or misuse.
A visitor should also know the shape of a consent form when one appears. Consent is asked in its own sentence with its own tick box, described in clear words and never hidden inside a page of legal type. Saying no, or taking back a yes later, never costs a visitor access to the public content of this site and never blocks a fair scoping conversation. The Company treats consent as something a free adult grants in full knowledge, not as a default the interface assumes.
Standard serving technology records a small set of operational signs when a page loads: internet protocol address, browser type, screen size, access time and the pages requested. The Company uses these signs to keep the site fast, to spot abuse and to count rough traffic so the bench knows which workstreams interest operations teams.
This site does not run marketing pixels and does not build cross site profiles of readers. Analytics, where present, are aggregated and short lived. The Company prefers fewer signals over more, because a modelling shop has no reason to chase visitor data it will never use to improve a solver.
The workbench and this website are directed at professionals who manage fleets, crews, rosters and operations. They are not directed at children, and the Company does not knowingly collect information from anyone under the age of sixteen.
If a parent or guardian believes a child submitted a contact form or any other information to the Company, they should write to note@yialgorithms.hair and ask for removal. The Company will delete the record promptly and confirm the deletion. No child marketing is ever used on this bench, now or in future.
The Company uses the information it holds for five connected purposes. To answer a problem note or enquiry. To scope, derive, test and tune a modelling engagement a client requests. To keep this website working and protected. To comply with laws that bind the Company. And to improve the robustness of the shared engineering practice, without ever disclosing a specific client project in identifiable form.
Contact details held for one purpose are not quietly turned to another. A scoping email is not treated as consent to run marketing campaigns, and the Company does not rent, sell or trade visitor contact lists with outside parties for their own promotion.
The Company relies on clear and honest reasons for each kind of processing. When a visitor sends a problem note, the Company processes that note because the visitor requested a service step, which is a legitimate contact. When a client signs an engagement, the Company processes operational data under the contract that engagement creates.
Some internal processing rests on a legitimate interest in keeping the bench secure and the site operating, weighed carefully against the modest privacy effect of the small records involved. For any resident whose law requires a specific legal ground, the Company will state the ground in the correspondence for that individual on request. Consent, where it is the applicable ground, is always separate, informed and easy to withdraw.
To run its work the Company relies on a small set of trusted vessels: a web host serving these pages, an email provider carrying correspondence, a file carrier moving documents a client approves, and occasionally a compute service where a very large derivation needs more cores than the bench keeps locally.
The Company keeps a current written list of these vessels and reviews it each quarter. A client may ask through note@yialgorithms.hair for the list relevant to its engagement. Before any new vessel receives project data, the Company checks its security posture and adds it to the reviewed list so the record stays honest.
The Company keeps different records for different lengths and says so plainly. A routine contact form note is kept only as long as the enquiry stays live, and is deleted when the thread closes unless it becomes a client file. Automatic serving logs are rolled over within a few weeks. Client model inputs are kept for the life of the engagement plus a brief archive period defined in the client agreement.
When a retention period ends, the Company deletes or depersonalizes the record rather than storing it forever on the chance it may help one day. Data held only to satisfy a tax or legal duty is minimised to what that duty truly needs.
Because the bench handles operational data that shapes a real floor, the Company keeps protections deliberately tight. Transmission to the site and to model endpoints runs over encrypted channels. File shares are access controlled per role, with least privilege the default. Credentials never sit in plain text, and the Company rotates keys on a schedule.
Backups run regularly and are tested by real restores, so a plan is always recoverable. Access by any person is logged and reviewed. No system is perfectly proof against every attacker, so the Company also trains its team to spot social attempts and to treat a pile of route data as the valuable asset it is.
Primary project records for YI ALGORITHMS LLC are kept in the United States, where the Company is registered at 4277 W Watchmen Way, Herriman - 84096-2759, United States (US). Supporting copies may sit with the reviewed vessels named above, most of which operate data centres in the United States and parts of the European Union.
The Company always knows where a copy lives. If a client requires records to remain within a specific region, the Company configures storage to that region and states the choice in the engagement notes rather than promising a thing it cannot enforce.
A visitor or an individual whose data the Company holds may ask to see a copy, to correct a mistaken detail, to delete a record, to limit how it is used, or to receive the record in a portable machine readable shape where the law grants that right. The Company answers such a request promptly and never penalises anyone for asking.
To exercise a right, write to note@yialgorithms.hair and identify the data at issue. The Company will confirm the hand belongs to the right person before releasing anything, using a reply to the known address as the check. Refusals are rare and always explained in plain words so the requester knows the reason and the next step.
Some clients operate across borders, and a routing or roster question may pull records that normally live in several regions. When the Company moves data across a border, it does so under a lawful mechanism and under the duty of the client agreement rather than by quiet convenience.
Transfers are logged so a client can see where a copy went. Where a transfer needs an explicit contract clause, the Company signs that clause before the bytes move. The Company takes the view that where a route reveals operational truth, the border that data crosses should be a deliberate, documented decision.
During a long engagement the Company may keep archive copies of model inputs, solver outputs and review notes so a result can be replayed if a client challenges it months later. Archives are sealed read only and access controlled.
Corrections are handled with the same care as the original build. When a client finds that a source feed carried a wrong zone or a wrong unit, the Company corrects the stored copy, marks the correction clearly, and re-runs any dependent prediction that the error could have poisoned. A silent edit that leaves an old number to confuse a later report is not how the bench works.
Operational data that a client brings to the bench is that client property. The Company treats it as confidential and never folds one client behaviour into another client model, because two fleets are never the same enough for that shortcut to be safe.
Synthetic test data the Company invents to stress a schedule is the Company own work and may be reused across projects in depersonalized form. Real client data, even scrubbed of names, stays inside that client sphere unless the client expressly agrees to share a general lesson with the field.
The Company reviews this policy as its work and the law evolve. When a change is material, the Company will post the updated text on this page with a refreshed date and, for clients under an agreement, will note the update in the ordinary engagement contact.
Material means a change to what data is kept, who may see it, or how long it stays, not a rewording of punctuation. A visitor who keeps using the site after a posted change accepts the newer text; a client is instead bound by the written agreement, which always leads where the two conflict.
This policy tells you what the Company does with information. The companion Terms of Service tell you the wider rules of using the site and sending the Company work. Both documents should be read together, because privacy habits mean little without the boundaries the terms set on use and liability.
The privacy page and the terms page share the same company name, the same registered address and the same contact. Neither page overrides the other; where a clause of one seems to clash with a clause of the other, the written client agreement settles the matter for that client.
Questions about this policy, about a right you wish to exercise, or about a data handling choice in a live engagement should be sent to the address nearest the matter. Email reaches the Company fastest at note@yialgorithms.hair. The telephone line is +12408395358.
The registered office for written correspondence is YI ALGORITHMS LLC, 4277 W Watchmen Way, Herriman - 84096-2759, United States (US). Letters sent there are handled during bench hours and answered through the address on the envelope so the sender can trust the reply is going somewhere real and controlled.