Top Tech Jobs & Startup Jobs in NYC, NY

Lead Replay’s company-wide security program as a hands-on individual contributor. Build and verify controls for application security, cloud systems, identity, authorization, sensitive data, AI workflows, vendors, workforce access, and human operations. Conduct threat modeling, architecture and code reviews, incident response, security testing, and secure tooling. Maintain SOC 2 evidence and customer trust controls while enabling product, data, machine learning, and platform teams to operate securely and independently.

Owns large-scale data projects for frontier AI labs from scoping through delivery. The role translates research needs into data specifications, quality requirements, metrics, and timelines; manages client relationships; coordinates partner companies and internal teams; removes blockers; and ensures high-quality execution. Success requires strong project ownership, analytical ability, communication skills, and comfort working with technical and research stakeholders in ambiguous, high-pressure environments.

Replay invites talented candidates who do not see a conventional role that fits to describe what they have built, what they would like to own, and how they could improve the company. The opportunity is open-ended and may involve creating a function or role aligned with Replay’s data-focused business and rapid growth.

Own a dedicated client portfolio through startup dissolution, providing frontline support, onboarding, process guidance, and operational execution. Responsibilities include dissolution filings, asset transfers, account closures, investor distributions, issue resolution, help-center content, and product improvements. The role requires strong organization, communication, judgment, and the ability to explain complex legal, tax, and operational processes to diverse stakeholders. This is an in-person, high-ownership role in New York City.

Design and operate backend systems for high-volume, multi-stage data pipelines that de-identify sensitive enterprise data. Own versioned contracts, lineage, state transitions, retries, replay, reconciliation, verification, migrations, and safe recovery. Diagnose production behavior across services and storage, improve correctness and throughput, and build auditable release controls. Collaborate with ML, data, platform, security, product, and customer-facing teams while using AI tools with deterministic safeguards and human review.

Build end-to-end product features across frontend, backend, data, testing, and observability. Own ambiguous customer problems from discovery through reliable production, improve system reliability and measurements, and integrate AI responsibly. Work with product, design, and customers to make complex workflows understandable, recoverable, and verifiable.

Lead and grow an engineering team responsible for enterprise data ingestion, de-identification, ML/evaluation, orchestration, and reliable operations. Hire, coach, define ownership, drive technical direction, review designs and PRs, debug incidents, prototype selectively, and establish lightweight practices for planning, releases, incidents, and quality while keeping hands-on involvement to enable team autonomy.

Build and operate a shared platform to enable product, data, and AI teams to ship reliable, secure, observable, and cost-aware systems. Create IaC modules, runtime templates, deployment workflows, telemetry, recovery and incident processes, self-service developer tooling, and cloud cost and security controls. Drive measurable reliability and reduce toil while partnering with Security and engineering teams.

Build and productionize ML systems for de-identification: improve NER, entity resolution, extraction, and document understanding. Drive error analysis, dataset creation, experiments, evaluation, safe rollouts, inference optimization, and active-learning/LLM-assisted pipelines to improve precision, recall, and utility in live pipelines.

Establish and run evaluation practices for de-identified enterprise data: build corpora, design experiments, quantify uncertainty, detect failure slices, and produce decision-grade evidence using Python, SQL, synthetic data, replay, and human/model judgments to guide model, pipeline, and release decisions.