You will turn long-running, manual, and tribal-knowledge processes into well-architected, tested, and observable Python services and frameworks that run against high-volume operational data on the order of hundreds of thousands of repair-order records per day flowing through our AWS and Snowflake ecosystem.
Most important: deep, demonstrable Python mastery and a track record of architecting complex Python systems (frameworks, SDKs, distributed processing engines, or plugin-based platforms). We will trade breadth
of data-platform tooling for depth in Python; Snowflake, Airflow, and AWS specifics can be learned on the job.
Key Responsibilities
1. Python Architecture & System Design
- Architect complex, modular Python systems layered and hexagonal (ports-and-adapters) designs, plugin and registry architectures, dependency injection, and clean separation between domain logic,
orchestration, and I/O.
- Build internal frameworks and SDKs that other engineers consume: reusable pipeline components, configuration-driven job engines, custom Airflow operators and DAG factories, and well-versioned shared libraries.
- Apply design patterns with judgment (strategy, factory, pipeline/chain-of-responsibility, observer,command) and advanced language features protocols and ABCs, generics, descriptors, metaclasses,decorators, and context managers where they reduce complexity rather than add it.
2. Concurrency, Performance & Scale
- Design concurrent and parallel workloads using asyncio, threading, multiprocessing, and concurrent.futures, with a clear understanding of the GIL, I/O-bound vs. CPU-bound trade-offs,backpressure, and cancellation.
- Stay current with the modern CPython runtime understand free-threaded (no-GIL) builds introduced in Python 3.13 and officially supported in 3.14, sub-interpreters, and the JIT, and judge when they are worth adopting versus when dependencies or single-threaded overhead make them premature.
- Profile and optimize hot paths with cProfile, py-spy, and memory profilers; use vectorized processing (Polars, pandas, NumPy, PyArrow), streaming/generator-based designs, and chunked I/O to process large datasets within bounded memory.
- Build resilient distributed processing idempotent jobs, retries with backoff, checkpointing,partitioned/batched execution, and graceful recovery for long-running workloads.
3. Services, APIs & Integrations
- Develop production services and APIs (FastAPI or similar) on async-first foundations, with strong request/response contracts using Pydantic v2, SQLAlchemy 2.x for relational access, and clients for third-party APIs, SFTP/FTP transfers, and cloud object storage.
- Integrate Python with the data platform Snowflake (Python connector, Snowpark), AWS (boto3, S3,SQS/SNS, Lambda, Step Functions), and Airflow through clean adapter layers rather than tightly coupled scripts.
4. Code Quality, Testing & Delivery
- Set the bar for Python engineering standards strict typing (mypy/pyright), linting and formatting (ruff/black), packaging (Poetry/uv, pyproject.toml), and semantic versioning of internal libraries.
- Own automated testing with pytest fixtures, parametrization, mocking at architectural seams,property-based testing (Hypothesis), and integration tests against real or containerized dependencies and embed data-quality checks directly into the code you ship.
- Drive CI/CD and observability reproducible builds, containerized deployments (Docker), structured logging, OpenTelemetry traces and metrics, and alerting so failures surface before users notice them.
5. AI-Enabled Engineering & LLM Integration
- Build agentic automation for data workflows that use LLM tool calling and multi-step orchestration to remove repetitive data-engineering tasks.
- Design and maintain Model Context Protocol (MCP) servers (e.g. FastMCP) that expose our data,pipelines, and internal tools to AI assistants through well-defined, secure, typed interfaces.
- Engineer for safety and reliability guardrails, structured outputs validated with Pydantic, evaluation harnesses, and awareness of prompt-injection risks in tool-using systems.
- Work fluently with AI coding assistants (Claude Code, GitHub Copilot, Cursor, or similar) while holding generated code to the same architectural, typing, and testing standards as hand-written code.
6. Technical Leadership & Collaboration
- Mentor engineers on Python design, idiomatic code, testing, and performance; lead rigorous,constructive code reviews.
- Partner with architects, data engineers, QA, and product to translate requirements and existing manual processes into maintainable, documented, production-grade software.
Required Qualifications
- Experience: 7+ years of professional software engineering, with at least 5 years building productionsystems primarily in Python.
- Python mastery (primary): Expert-level command of the language and its internals data model, typing system, iterators/generators, decorators, descriptors, metaclasses, context managers, memory model and the GIL. You can explain why, not just how.
- Complex Python architecture: Proven experience designing and owning large, multi-module Python codebases or frameworks used by other engineers not only standalone scripts or notebooks.
- Concurrency & performance: Hands-on production experience with asyncio and/or multiprocessing, and a track record of profiling and measurably improving the speed or memory footprint of Python systems.
- Testing discipline: Deep experience with pytest and test design for complex systems, including mocking strategy, integration testing, and test-driven development.
- Data handling: Strong experience processing large datasets in Python (Polars, pandas, PyArrow, DuckDB, or similar), dataframe and schema validation (Pandera or Pydantic), and solid working SQL.
- Modern Python ecosystem: Current with Python 3.12+ and modern tooling uv, ruff, pyright/mypy,Pydantic v2, async-native libraries (httpx, asyncpg/aiohttp) and able to evaluate new runtime features critically.
- Engineering practices: Git-based workflows, CI/CD, packaging and dependency management,containerization, and strong documentation habits (design docs, READMEs, runbooks).
Preferred / Nice-to-Have
- Data platform exposure: Experience with Snowflake, Airflow (including custom operators/DAG factories),dbt, or Databricks/PySpark. Helpful, but not required for the right Python engineer.
- AWS & event-driven systems: Hands-on use of S3, Lambda, SQS/SNS, Step Functions, or ECS from Python, including idempotent consumers for Kafka, Kinesis, or SQS.
- Performance extensions: Experience with Cython, Rust bindings (PyO3), or Numba, or with running workloads on the free-threaded CPython build.
- LLM & agent frameworks: Hands-on experience with MCP server development, LangGraph, Pydantic AI,or similar agent frameworks, and with LLM APIs such as Anthropic, OpenAI, or AWS Bedrock.
- Automotive domain: Familiarity with repair-order, dealership fixed-operations, or DMS data is a plus.
Shortlisted candidates should expect a technical interview that includes a live architecture discussion and an
in-depth review of real Python code