The posting
Be at the Forefront of the Agentic AI Revolution
At Skan AI, we are pioneering the context engine for human and agentic execution, bringing context from enterprise operators, systems, and processes to power how the world's largest organizations execute their most complex, mission-critical work.
Why Skan AI
We're in hyper-growth mode at exactly the right moment in history. As enterprises race to adopt agentic AI, we're uniquely positioned to deliver the clear signal they desperately need: a platform that trains and grounds AI Agents in trillions of real execution signals, enabling reliable, compliant automation of their most complex processes.
Backed by Dell Technologies Capital and other leading investors, we're the only company that can bridge the gap between AI's promise and enterprise reality, making us perfectly positioned to define the agentic era for modern enterprises.
Our diverse, collaborative team of 250+ innovators is solving category-defining challenges at the intersection of AI, process intelligence, and enterprise work. Diverse perspectives fuel breakthrough thinking, cross-functional collaboration is the norm, and our work directly transforms how Fortune 500 companies operate. We are shaping the future of work itself.
About the Role
The Context Graph team develops the systems that turn large volumes of observed enterprise work into structured process knowledge. Much of that work begins as rapidly evolving research code, experimental harnesses, model-backed pipelines, and new algorithmic components. This role is responsible for turning those systems into production software that can be trusted at scale.
As a Software Engineer on the Context Graph team, you will own the engineering boundary between research and production. You will take systems that work in controlled development settings and make them reliable under real workloads, real data volumes, real deployment constraints, and real failure conditions.
That means much more than writing services. You will design production interfaces around research components, build scalable execution paths, define rigorous testing and validation strategies, trace regressions across complex data pipelines, and make sure changes that improve one part of the system do not quietly degrade another.
This responsibility is becoming more important as software is increasingly written with agentic coding systems. Code can now be produced faster than it can be deeply understood. We therefore care disproportionately about engineers who bring judgment, code review discipline, testing rigor, systems understanding, and an instinct for finding where seemingly correct code will fail under pressure.
The systems themselves are distributed and heterogeneous. Data may originate from observation software running inside an enterprise environment, move through gateway and processing layers, pass through asynchronous multi-stage pipelines, interact with model-backed services, and ultimately execute across cloud, hybrid-cloud, or fully on-premise deployments.
Your job is to make that entire path boring in the best possible way: predictable, measurable, testable, scalable, and difficult to break.
Responsibilities
Productionize research and agentic systems
- Take research prototypes, algorithmic harnesses, and model-backed pipeline components and turn them into reliable production services.
- Define clear interfaces, contracts, failure behavior, and operational boundaries around rapidly evolving research code.
- Build integration and regression suites that allow research teams to iterate quickly without destabilizing existing production behavior.
- Review and harden code produced across both human and agentic development workflows, with particular attention to hidden assumptions, incomplete failure handling, and unintended side effects.
- Establish patterns for moving experimental capabilities into the existing product line without creating fragile parallel systems.
Build and scale distributed services
- Design, build, and operate production Python microservices.
- Build asynchronous, message-driven services on RabbitMQ or equivalent systems, including retry semantics, idempotency, backpressure, and failure recovery across multi-stage pipelines.
- Develop and maintain workflows on our in-house orchestration engine, Temporal, and Airflow.
- Design and evolve data layers across PostgreSQL or MySQL, MinIO object storage, Redis, and high-volume intermediate artifacts.
- Build and scale the service layer around LLM and agentic pipelines, including throughput, concurrency, caching, latency, and cost.
- Reason about the behavior of services as workloads move from development-scale runs to enterprise-scale execution.
Own correctness and regression resistance
- Build rigorous unit, integration, system, replay, load, and regression testing for complex pipeline behavior.
- Design scalable test harnesses that exercise research and production systems across realistic workloads, failure modes, and deployment configurations.
- Develop performance benchmarks for throughput, latency, resource utilization, and queue behavior, and identify limits before customers do.
- Trace regressions across long data paths spanning ingestion, gateways, orchestration, services, model calls, storage, and output generation.
- Establish validation gates that prevent changes from reaching production when correctness, performance, or reliability has degraded.
- Make failures reproducible enough that complex production issues can be diagnosed rather than merely observed.
Build for heterogeneous deployment environments
- Build and maintain CI/CD pipelines covering build, test, artifact publishing, validation, and deployment.
- Write and maintain Dockerfiles, container images, Kubernetes manifests, Helm charts, and versioning strategies.
- Manage infrastructure as code with Terraform across GCP, Azure, and supported enterprise environments.
- Design services that behave consistently across cloud, hybrid-cloud, and on-premise deployments.
- Own environment configuration, secrets management, releases, rollbacks, and deployment troubleshooting.
- Account explicitly for the constraints that emerge in customer-controlled environments, including network boundaries, resource limits, service dependencies, and restricted infrastructure.
Operate what you build
- Instrument services with high-quality logs, metrics, and traces, and build the dashboards and alerts needed to understand production behavior.
- Use observability to diagnose system behavior across asynchronous and distributed execution paths.
- Support GPU-hosted and other resource-intensive workloads, including scheduling, resource allocation, and cost control.
- Participate in production support and on-call for the services your team owns.
- Treat operability as a design requirement rather than something added after deployment.
Collaboration
- Work directly with applied scientists and AI engineers to understand the intended behavior of new algorithms before productionizing them.
- Work with product engineers and platform teams to integrate new Context Graph capabilities into the existing product without creating regressions.
- Participate deeply in technical design, code review, testing strategy, and architecture discussions.
- Act as a quality bar for systems entering production, especially when code spans multiple teams or has been developed rapidly through agentic coding workflows.
Requirements
- We screen for comparable experience and strong fundamentals, not exact tool matches. Equivalent technologies are welcome.
- Ambitious and Hungry. “Far and away the best prize that life offers is the chance to work hard at work worth doing.” - Theodore Roosevelt
- Production Backend Experience. 2-5 years building and operating production backend systems in distributed or service-oriented environments.
- Strong Python. You write clear, tested, production-quality Python and can quickly understand unfamiliar code well enough to challenge its assumptions, failure behavior, and scalability.
- System Design Fundamentals. You are comfortable reasoning about horizontal and vertical scaling, autoscaling, load balancing, capacity planning, caching, statelessness, concurrency, and throughput versus latency tradeoffs.
- Distributed Systems Fundamentals. You understand asynchronous and event-driven systems, queue-based load leveling, idempotency, retries, timeouts, backpressure, graceful degradation, and how failures propagate across service boundaries.
- Testing and Software Quality. You have strong instincts around unit, integration, system, replay, performance, and regression testing. You know that high coverage is not the same as high confidence and can identify which invariants and failure modes actually need protection.
- Performance Engineering. You can reason about bottlenecks across CPU, memory, storage, networking, queues, database access, and model-backed services, and design tests that reveal scaling limits before they become production incidents.
- Message Queues in Production. Experience with RabbitMQ, Kafka, SQS, Pub/Sub, or equivalent event-driven infrastructure, including consumer lag, ordering, retry behavior, and failure recovery.
- Data Layer Fundamentals. Strong understanding of relational schema design, query performance, transactions, connection management, caching, and object-storage-backed data flows.
- Containers and Orchestration. You understand how services are packaged, scheduled, scaled, isolated, given resource limits, and operated across different deployment environments.
- Deployment and Infrastructure. Experience with CI/CD and infrastructure as code, and an understanding of the differences between cloud-native, hybrid-cloud, and on-premise production environments.
- Operability. You can debug live systems using logs, metrics, traces, dashboards, and raw system behavior, and are comfortable reasoning through failures that span multiple services.
- Engineering Judgment. You are skeptical of code that merely works once. You look for hidden coupling, untested assumptions, scalability limits, and ways a change can create regressions elsewhere in the system.
Preferred
Direct experience with any of the following is valuable:
- Workflow orchestration engines such as Temporal or Airflow, or experience building similar systems.
- High-throughput APIs, gateways, rate limiting, connection pooling, and latency tuning.
- Object storage as a primary data plane, including S3, GCS, or MinIO.
- LLM application infrastructure including model gateways, prompt caching, token and cost management, and model observability.
- Large-scale data pipelines involving images, event streams, or unstructured artifacts.
- Observability systems such as Grafana, Loki, Tempo, Prometheus, or Mimir.
- Multi-cloud or hybrid-cloud deployments across GCP, Azure, or enterprise-controlled environments.
- On-premise software deployment and operating systems under constrained infrastructure or network conditions.
- GPU workload scheduling and cost optimization.
- Experience building performance, replay, or regression harnesses for complex distributed systems.
You Might Be a Great Fit If You
- Are uncomfortable merging code you do not deeply trust, even when it appears to work.
- Think about failure modes, invariants, and ten-times scale before a service reaches production.
- Like being the engineer who can take a fast-moving research system and make it safe for the rest of the company to depend on.
- Find a subtle regression across five services more interesting than a straightforward greenfield API.
- Care about tests that prove behavior rather than tests that merely increase coverage.
- Enjoy owning a service through design, deployment, observability, production debugging, and on-call.
- Are interested in the new engineering problem created by agentic coding: software can be generated extremely quickly, so judgment, validation, and systems rigor become more valuable, not less.
Skan AI is an equal opportunity employer committed to building a diverse, inclusive, and respectful workplace around the world. We do not discriminate based on race, color, religion or belief, sex (including pregnancy, sexual orientation, gender identity, or gender expression), national origin, ancestry, age, disability, medical condition, genetic information, marital or family status, military or veteran status, or any other characteristic protected by applicable laws in the locations where we operate.
We welcome people from all backgrounds and provide reasonable accommodations throughout the hiring process
