About the Role :

You will own backend services that process large, complex mortgage documents at scale, designing the pipelines, integrating LLMs, and keeping the whole thing fast, accurate, and cost-effective in production. This is a hands-on senior role with real ownership: you will make architecture calls, drive them to a decision, and set the quality bar for engineers working alongside you.

Responsibilities :

- Build high-throughput document pipelines: Asynchronous, parallel processing with batching, retries, backpressure, and idempotent reprocessing for large documents and large data volumes.

- Work with the latest LLMs: Build against current LLM API timeouts, token and cost controls, response validation, and automatic failover between providers.

- Engineer for resilience: Circuit breakers, fallback chains, and graceful degradation so a single upstream failure never takes the pipeline down.

- Debug production: Investigate extraction failures, latency spikes, and cost regressions using dashboards and metrics, and fix the root cause.

- Raise the bar on the team: Review PRs, break features into clear tasks, unblock junior engineers, and write design docs that hold up to scrutiny across time zones.

Requirements :

- Bachelor's degree in Computer Science, Information Technology, or a related field.

- 4+ years of hands-on Java development, including at least 2 years owning a production service.

- Excellent command of Core Java, with real depth in multithreading and concurrent architecture design.

- Strong with Spring Boot and Spring Data; comfortable in microservices and REST architecture.

- Strong working knowledge of NoSQL DB (preferred: MongoDB); familiar with relational databases (MySQL or PostgreSQL).

- Comfortable with Docker and Kubernetes, and at least one cloud platform (AWS, GCP, or Azure).

- Strong debugging and problem-solving skills with a solid command of data structures.

- Reasons about trade-offs in cost, latency, and accuracy, not just correctness.

- Has informally mentored or guided junior engineers.

- Operates independently with minimal supervision, and proactively flags risks.

- Excellent written communication; much of the collaboration is asynchronous and across time zones.

Good to Have :

- Exposure to LLM or GenAI application development, prompt iteration, evaluation, or accuracy measurement.

- Working knowledge of Python.

- Experience with streaming queues like Kafka or RabbitMQ.

- Background in Document AI, OCR, or the mortgage/fintech domain.

- Familiarity with observability tooling (Grafana, Metabase, ELK).