Skip to content
IDT Partners
IDT Partners
New York
New York • Senior engineers • Selective engagements

Serious engineering for products that matter.

IDT Partners designs and builds products, applied AI, and the platforms they run on. We stay with the work from architecture through delivery.

What we build
Applied AI
Search, assistants, and workflows built on your own data.
Distributed systems
Services that scale across teams and traffic.
Data platforms
Pipelines and storage built for the product they serve.
Selected experience
Adidas American Express Bank of America Barclays Disney The New York Times UBS

Services

We work with product and platform teams that need senior engineers for architecture, applied AI, and delivery.

Applied AI

Retrieval, classification, and assistants for support and operations, plus multi-step workflows with human approval. The models behind them can be replaced without rework.

Distributed systems

Event-driven services, queues, and APIs with explicit contracts and clear ownership of data.

Data platforms

Ingestion, pipelines, storage, and access control, designed alongside the product that depends on them.

Product engineering

Fast, accessible web and mobile applications, built by the engineers who design the services behind them.

Integration

Connections to the systems you already run, such as identity, CRM, and internal tools, without rewriting them.

Modernization

Move off a legacy platform in stages while the product keeps running.

Approach

1) Discovery
Define the work
  • Scope, constraints, and measures of success
  • Where AI helps and where it does not
  • Target architecture and a plan for the first release
2) Delivery
Build to production
  • Working software reviewed every week
  • Automated tests and code review on every change
  • Security and performance reviewed before launch
3) Launch
Transfer ownership
  • Staged rollout with monitoring in place
  • Runbooks, dashboards, and documentation
  • Handover to the engineers who will run it

Engineering standards

Evaluation

Models are tested against fixed evaluation sets before each release and monitored afterward. People review the output where the stakes require it.

Reliability

Timeouts, retries, and idempotent operations are designed in, and every request can be traced across services.

Security

Least-privilege access and audit logging throughout, with data handling matched to your risk profile.

Cost and performance

Latency and spend are measured in production, then tuned with caching, batching, and model routing.

Contact

Write to us about the work. We will reply with an assessment of fit and timing.

Email

Responses typically within one business day.

NYC Office
253 W 35th St
New York, NY 10001

Meetings in New York, by appointment.