As an Engineering Manager at Blend, you will lead and grow a team of engineers building Blend's intelligent origination products including Blend Autopilot, our AI agent that automates document review, compliance checking, and borrower follow-ups during loan origination, taking documents from upload to reviewed in 15–25 seconds. This role offers a unique opportunity to shape one of the most AI-forward teams in Fintech: we ship agentic features rapidly, lean hard on modern AI tooling at every layer of the stack, and are looking for a leader who will continue pushing that pace of adoption: both in the products we deliver and in how we build them.

Key Responsibilities:

• Guide architectural decisions, ensuring the delivery of scalable, robust, and maintainable solutions aligned with Blend's goals.
• Hire, mentor, and develop a team of high-performing engineers, fostering a culture of ownership, innovation, and accountability.
• Collaborate with product managers, design teams, and other stakeholders to align technical execution with business objectives.
• Drive the end-to-end development lifecycle, including design, implementation, testing, deployment, and production maintenance.
• Play a key role in connecting global engineering teams, ensuring alignment across teams, tools, and processes.
• Conduct regular code reviews, advocate for engineering best practices, and encourage technical excellence.
• Champion an AI-forward engineering culture with rapid adoption of new foundation-model capabilities, agentic patterns, and AI-assisted developer workflows, and keep raising the bar on how the team builds with AI.

Preferred Qualifications:

• Bachelor's degree in Computer Science, Engineering, or a related field.
• 6+ years of software engineering experience, with at least 2 years in a leadership or managerial role; proven ability to hire, mentor, and grow engineers.
• Excellent communication skills with the ability to work independently and collaboratively in a fast-paced, globally distributed environment.
• Deep Python expertise (3.11+), with strong command of async/await, asyncio, type hints, and Pydantic v2. Our codebase is ~98% Python and async-first.
• Hands-on experience building production agentic systems: multi-agent orchestration, tool use, structured outputs, context management, and hallucination mitigation. Experience with agent dev kits (ADK, DSPy, LangGraph, or equivalent) and durable workflow engines such as Temporal, Airflow, or Step Functions.
• Foundation-model fluency across providers: Google, Anthropic, or OpenAI. Including model selection, prompt caching, config sweeps, and cost/latency tradeoffs. Familiarity with multi-provider abstractions (e.g. LiteLLM) is a strong plus.
• Prompt engineering and agent context management: hands-on experience designing system prompts and tool schemas, managing context windows (prompt caching, summarization, retrieval injection), and shaping model behavior through few-shot examples, structured outputs, or fine-tuning. Comfortable treating prompts as first-class engineering artifacts backed by evals and version control.
• Agent-assisted development workflows: fluency with Claude Code (or comparable AI coding agents), git worktrees for parallel agent sessions, and AI-assisted code review. We expect everyone to use AI tooling daily, and we want a leader who will continue pushing the team's adoption forward.
• Retrieval-Augmented Generation (RAG) experience with vector databases (Qdrant, Pinecone, or similar), embedding models (sentence-transformers), and ideally multimodal retrieval (ColPali, FAISS) over enterprise documents.
• Eval-driven development of LLM features: frameworks like Braintrust, LangSmith, or MLflow with PII masking, scenario design, and CI-enforced quality baselines.
• PII / regulated-data sensibility: experience with redaction tooling (Presidio or similar), structured logging, and audit-grade observability. Comfort working in a SOC 2 / financial-services environment.
• Cloud + container ops: AWS (EKS, RDS, S3) and Kubernetes; comfortable with Docker, GitHub Actions, and modern Python tooling (uv, ruff, pytest-asyncio, pre-commit, import-linter).

Bonus: exposure to mortgage / loan-origination workflows (URLA, AUS, document validation, follow-ups). Helpful but not required; we hire strong engineers and teach the domain.