Job Description

We are looking for an experienced Lead AI Engineer to work on an exciting project - Tax Notice Navigator which automates government tax notice workflows end-to-end: document ingestion, OCR, classification, routing, response generation, and case resolution. As the first AI engineer on the product, you'll own architecture, delivery, and technical direction across all AI components.

What You'll Do

Design and deliver the platform's full AI stack: OCR ? parsing ? classification ? LLM orchestration ? rules ? storage. Implement the initial core use cases: Rules automation & assignment Document extraction & classification Templated response generation AI-assisted investigations Continuous prompt refinement Build scalable, auditable workflows that reduce manual work and improve accuracy. Collaborate with product, engineering, and tax SMEs to define the roadmap and technical standards. Establish evaluation, monitoring, and CI/CD for AI pipelines.

Key Responsibilities

AI Platform & Architecture

Integrate OCR/Document AI tools and normalize outputs. Build LLM orchestration and multi-agent workflows with guardrails, retries, and state. Implement a simple rules engine for assignment, categorization, and automated actions.

Document & Data Pipeline

Build an OCR/document pipeline with QC, layout parsing, and entity extraction. Design schemas for documents, extracted fields, and audit logs.

Response Generation & Investigation

Create a templated response library tied to extracted data. Generate safe, compliant AI drafts with human-in-the-loop review. Implement an investigation assistant that gathers context and recommends next steps.

Prompt Engineering & Evaluation

Maintain a versioned prompt library with regression tests and offline evaluation datasets. Track metrics for accuracy, latency, safety, and response fidelity.

MLOps & Security

Containerize and deploy services (Docker/Kubernetes). Set up monitoring for drift, cost, and quality (Prometheus, LangSmith). Ensure strong PII protection, RBAC, audit trails, and compliance.

Minimum Qualifications

6+ years in AI/ML or NLP; 1.5+ years building

production LLM applications

. Expert Python and strong experience with LLM orchestration (LangGraph or similar). Hands-on with agent frameworks (e.g., CrewAI). Document AI/OCR experience (Textract, Form Recognizer, Document AI, Tesseract). Practical RAG, vector DBs, and structured output validation. Experience with rules engines or DSLs. Strong prompt engineering, evaluation, and safety practices. Production APIs (FastAPI/Flask), Docker, CI/CD, cloud (AWS/Azure). Deep knowledge of PII protection, RBAC, and audit logging.

Preferred

Event-driven systems (Service Bus), Celery/Redis workers. LLM cost and latency tuning.

Tech Stack

Python

LangGraph

CrewAI

OpenAI/Anthropic/Azure

Textract/Form Recognizer

FastAPI

Celery

Redis

Postgres + pgvector

S3/Blob

Prometheus

LangSmith

Docker

Kubernetes

Terraform

AWS/Azure

What do we offer you?

Attractive salary Large freedom and real influence No unhealthy competition, team approach to meeting challenges Remote-first, flexible working culture Company apartments in cool cities across Europe: work and enjoy a memorable getaway

About Us

We are a software house with 18 years of experience and a global portfolio of projects. We help businesses modernize, scale, and innovate through custom software solutions -- always with a focus on flexibility and quality. Our team embraces unconventional ideas and new technologies, delivering solutions that drive real impact. If you value professionalism, creativity, and a strong engineering culture, you'll feel at home here.

Job Type: Full-time

Pay: 45,000.00-102,000.00 per year

Benefits:

Work from home
Experience:

AI / ML: 5 years (required)
Work Location: Remote