Staff Infrastructure Engineer

## About Alka Intelligence We are an asset management startup building a proprietary reasoning engine to automate fundamental analysis. Our system is capable of ingesting vast, unstructured market datasets and transforming them into clear, risk-weighted investment mandates. Our engine is designed to show its work, creating a fully auditable record of why a decision was made. We are looking for builders to help us bridge the gap between stochastic AI models and the rigorous reliability required to manage fiduciary capital. Locations: San Francisco and Dallas ## Role Overview As an Infrastructure Engineer, you will own the stack beneath our application layer including platform, cloud, reliability, data, and compliance. You will design and operate the infrastructure that runs our agents in production such as cloud architecture, container orchestration, CI/CD, and observability. You will own our data infrastructure end-to-end which includes ingestion pipelines, third-party market data integrations, warehousing, and the audit-grade trace storage that makes our "show your work" thesis defensible to investors and regulators. You will be responsible for incident response and building the paved road that lets our engineers ship quickly without breaking production. You will own considerable portions of our infrastructure and must be comfortable working in a rapidly iterative environment. We're looking for engineers who treat infrastructure as a product, automate aggressively, and understand that at a fintech, reliability and auditability are features and not afterthoughts. ## What You Have * Passion for owning infrastructure end-to-end in ambiguous environments. You automate repetitive work, make quick decisions, and deliver on accelerated deadlines. * Proficiency in Python or Go and deep experience architecting cloud infrastructure on AWS, GCP, or Azure, containerization (Docker, Kubernetes), infrastructure-as-code (Terraform, Pulumi), CI/CD, IAM, secrets management, databases, queues, and caching. * Experience building and operating ingestion pipelines and data platforms, orchestrators (Airflow, Dagster, Prefect), search and retrieval systems (Elasticsearch or OpenSearch), knowledge graphs (Neo4j or Neptune), and reliable patterns for integrating third-party data APIs. Hands-on experience with DuckDB or ClickHouse is a strong plus. We're greenfield on the data warehouse and you'll help us choose and stand it up. * Experience operating LLM-backed systems in production, managing provider integrations (Anthropic, OpenAI), cost and rate limit controls, and observability for non-deterministic workloads. * Prior ownership of production systems end-to-end including observability, incident response, and the security and compliance hygiene (SOC 2 or similar) that fintech requires. * Experience hosting LLMs and other deep learning models is a bonus. We don't host models right now, but that may change. * You don't need to be a finance expert, but you must have an interest in finance and investing. We only hire candidates that are passionate about our field. * 5+ years of professional infrastructure, platform, or SRE experience, with a preferable tour at an early-stage startup or as an early infrastructure hire. * Candidates must be currently residing in the US with valid work authorization or OPT status. H-1B sponsorship is evaluated on a case-by-case basis. ## Short Answer Questions * **Production Incident:** Please describe the most complex production incident you've owned end-to-end. Include the system involved, how you diagnosed it, what you did to resolve it, and what you changed afterward to prevent recurrence. Feel free to include links and keep your response below 300 words. * **Infra Tradeoff:** Tell us about a time you made a non-obvious infrastructure decision on AWS (or another cloud) that involved a real tradeoff such as cost vs. performance, build vs. buy, simplicity vs. flexibility. What were the alternatives, what did you choose, and what would you do differently now? Keep your response below 300 words.