James Westover

Senior Python Developer — 12+ Years Production Python | FastAPI, Django, Flask, Tornado, AsyncIO, Distributed Systems, AI/ML

Profile

Senior Python developer with 12+ years of production Python across the full version range (2.4 through 3.13) and the full framework spectrum — FastAPI, Django, Flask, Tornado — at companies ranging from high-growth startups to global-scale platforms. Deep expertise designing and shipping production REST APIs, event-driven microservices, and data-intensive backend systems. Proven track record on the hardest Python migration and scalability challenges: led company-wide Python 2→3 migrations at both Uber (90%+ of production services) and Postmates (1M+ LOC monorepo, 3 concurrent Django versions), built Cassandra infrastructure at 2–10M QPS, and architected real-time streaming pipelines processing 250M+ events per day. Currently building production FastAPI services with ONNX Runtime AI inference, PostgreSQL/pgvector semantic search, and modern packaging (PDM) through a Berlin-based consultancy. Experienced mentor, tech lead, and engineering manager who has grown teams from 2 to 15 engineers. Available for fully remote contract or permanent roles.

Senior Python Developer — Common Requirements Mapped to Experience:

Python (12+ years, all major frameworks): Production Python 2.4 through 3.13 across every major web framework — FastAPI (current, production inference APIs), Django (URBN product catalog + Postmates monorepo), Flask (inventory tooling at HERE), Tornado (HERE fleet APIs + Uber's tchannel-python RPC library). Fluent in asyncio, Celery, SQLAlchemy, Pydantic v2, Alembic, Click, structlog, and the broader stdlib.
REST API design & microservices: Designed production REST APIs at every role. Architected microservice ecosystems at Uber scale (~10,000 nodes per DC). Experience with gRPC, Apache Thrift, Protobuf, and Envoy service mesh. API versioning, auth patterns (token-based, GDPR-compliant), and OpenAPI documentation.
Testing & quality (pytest, tox, CI): Comprehensive pytest suites with fixtures, parametrize, and mocking. Tox for multi-version compatibility. Delivered 4x CI build time improvement at Postmates by distributing test execution across Kubernetes pods (40–50 min → 10–15 min). GitHub Actions and GitLab CI pipeline ownership.
Packaging & dependency management: Hands-on with PDM (current), pip, virtualenv, poetry, pipx, and pyenv. Managed Python dependency graphs spanning hundreds of microservices at Uber. PDM + uv resolver for fast, reproducible installs in current production systems.
Databases — relational & NoSQL: PostgreSQL (PostGIS, pgvector, pgpool, PgBouncer, Alembic migrations, query optimization), MySQL, Oracle. Cassandra (multi-region, 2–10M QPS, QUORUM/LOCAL_QUORUM/ONE consistency tuning, tombstone management). Redis (caching, pub/sub, HyperLogLog/MinHash probabilistic structures). Elasticsearch, Amazon Redshift, HBase, Google Bigtable.
AI/ML Python stack: Currently deploying FastAPI + ONNX Runtime inference services in production (sub-100ms object detection and NSFW classification). PostgreSQL with pgvector for semantic search and knowledge graph. Experience with PyTorch, TensorFlow, Scikit-learn, NumPy, SciPy, Pandas, OpenCV. AWS SageMaker for managed inference design (Transit Authority bid). Bayesian sports modelling with PyMC5 and Polars.
Cloud platforms (AWS, GCP, Azure): AWS (ECS, EC2, S3, Lambda, SageMaker, CloudFormation, CloudWatch, IAM, Kinesis, Redshift, Athena, EMR). GCP (Bigtable, Compute Engine). Azure (VMs, Storage). Hybrid cloud and on-prem deployments. Cloudflare edge (Tunnels, Pages, Workers, DNS). Tailscale zero-trust networking.
Docker & Kubernetes: Docker in production at Uber, Postmates, URBN, and personal projects. Kubernetes for distributed test execution (Postmates), service mesh integration (Envoy), and container-based CI/CD. Ansible for configuration management and immutable deployment patterns.
Event-driven & messaging: Apache Kafka (Confluent Platform, event sourcing, consumer groups). RabbitMQ (Sonobi: 250M+ daily impressions). Apache Storm (Python bolts, lambda architecture at HERE). Celery (URBN task processing). Amazon Kinesis. Built two custom workflow orchestration systems (Warehouser at Sonobi, Weatherman at HERE).
Python migrations & legacy modernization: Led Python 2→3 migration at Uber (90%+ of production services; mapped full dependency graph, targeted top 200 repos, achieved 93% conversion). Led same migration at Postmates (1M+ LOC monorepo; 3 coexisting Django versions; bytes-vs-unicode incompatibilities in performance-critical paths). Major contributor to tchannel-python (Tornado-based Thrift RPC) Python 3 port.
Observability & reliability: Prometheus, Grafana, Datadog, ELK Stack, OpenTelemetry, structlog. On-call rotations and postmortems at Uber and Postmates. Incident command for GPS millisecond offset bug affecting millions of devices globally. Drove platform reliability from ~80% to 97%+ at Uber Location Platform.
Security & compliance: GDPR-compliant location data architecture at Uber. Data deletion workflows across distributed databases. ISO 9001 certification for safety-critical mapping software at HERE. Privacy-by-design patterns in current AI inference APIs (images never persisted post-analysis). AWS KMS, GitHub Secrets, GCP Secret Manager.

Selected Experience

Westover Labs / Neher Data Services — Berlin, Germany

Founder (Westover Labs) & CTO (Neher Data Services) · 2021 – Present

Build and ship production FastAPI services in Python 3.13 with ONNX Runtime AI inference, Pydantic v2 request/response validation, async endpoint design, and comprehensive pytest suites — deployed via Ansible and GitHub Actions CI/CD pipelines.
Architect and maintain PostgreSQL-backed knowledge graph service with pgvector for semantic similarity search, Alembic schema migrations, and dual REST + MCP API surfaces — all managed with PDM for reproducible dependency locking.
Build token authentication and subscription management APIs: Apple IAP validation, 24-hour token issuance (SQLite-backed), and per-tier rate limiting — privacy-first architecture where user images are never persisted post-analysis.
Implement Bayesian sports modelling system using PyMC5, Polars, and DuckDB for probabilistic DFS lineup optimization — full data pipeline from ingestion to probabilistic inference.
Design event-driven automation systems with browser orchestration (Playwright), message queues, and webhook-based workflows; structured logging via structlog across all services.
Pioneer AI-augmented development methodology — orchestrating 12 specialized AI agent personas to deliver team-level Python development throughput as a solo engineer.
German-registered company (UG) operating fully remotely across EU and international clients.

Uber — San Francisco, CA / Remote

Staff Software Engineer · February 2021 – October 2022

Senior technical contributor on Uber's navigation platform — one of the company's largest distributed Python + Java services (~10,000 nodes per data center), serving all routing and mapping globally across 30M+ driver devices.
Enhanced Python-based simulation tools scaling test capacity from 500 to 10,000 trips per run using multiprocessing parallelization — critical for testing off-route detection algorithms at scale.
Led Postmates acquisition technology integration: maintained business continuity for legacy systems while designing GDPR-compliant data deletion strategies across distributed databases using Go-based microservices.
Led critical incident response for GPS millisecond offset bug corrupting location data across millions of devices — diagnosed root cause, coordinated cross-team remediation.

Senior Software Engineer — Python Platform (Office of CTO) · January 2019 – June 2020

Founding member of dedicated platform team for Uber's enterprise-wide Python 2→3 migration — spanning hundreds of teams and 90%+ of production services.
Mapped entire Python dependency structure across Uber's microservice ecosystem; built purpose-built automation tooling to target top 200 repositories and achieve 93% service conversion.
Major contributor to tchannel-python — Tornado-based Thrift RPC transport library (Python 2→3 migration of core transport layer). Open-source, production-critical.
Became de facto Python support for Uber's data science organization — maintained notebook infrastructure, supported distributed query systems, debugged internal Airflow fork.
Provided mentorship and office hours company-wide for Python debugging, performance tuning, and version compatibility across engineering.

Software Engineering Manager — Location Platform · February 2018 – January 2019

Founded and built Uber's Location Platform team from scratch — designed GDPR-compliant Python-based location data pipeline for global fleet with Python ingestion services and Kafka-based event streaming.
Drove platform reliability from ~80% to 97%+ global data ingestion success, supporting Uber's RideCheck safety initiative.
Owned architecture for globally distributed Cassandra infrastructure at 2–10M QPS — consistency model tuning (QUORUM, LOCAL_QUORUM, ONE), tombstone management, multi-region capacity planning.
Directed anti-fraud program identifying GPS location spoofing — Python-based detection tooling delivering $30M+ in annual savings.

Postmates Inc. — San Francisco, CA

Senior Software Engineer 2 — Core Infrastructure · June 2020 – February 2021

Led monorepo-wide Python 2→3 migration across 1M+ lines of code serving 100 engineers — navigating 3 coexisting Django versions, deep bytes-vs-unicode incompatibilities, and performance-critical code paths.
Delivered 4x CI build time improvement by orchestrating distributed pytest execution across Kubernetes pods — transforming 40–50 minute feedback loops into 10–15 minutes.
Partnered with infrastructure team on Envoy service mesh and OpenTelemetry distributed tracing adoption — improving observability across Python microservices.
Participated in weekly war room calls during peak-traffic scaling incidents — real-time incident response alongside infrastructure leadership.
Worked with PostgreSQL (pgpool, PgBouncer) for connection pooling and query optimization across the monorepo's database layer.

HERE Technologies — Berlin, Germany

Software Engineering Manager → Senior Software Architect · April 2015 – February 2018

Built engineering organization delivering 8+ concurrent programs — scaled team from 2 to 15 engineers structured across backend (Python), frontend, mobile, DevOps, and embedded subteams.
Designed Tornado-based backend APIs with SQLAlchemy + PostgreSQL/PostGIS and Flask tooling for fleet management — serving real-time GPS, IMU, and camera telemetry for autonomous driving mapping vehicles.
Architected lambda data architecture using Apache Storm (Python bolts), Kafka, Elasticsearch, and PostgreSQL/PostGIS for vehicle fleet ingestion processing 250M+ GPS events/day and real-time analytics.
Built "Weatherman" Python orchestration system for dynamic map reprocessing workflows and QA — event-driven batch processing with Kafka-based coordination and scheduling.
Built Flask-based inventory management system using QR code scanning for real-time hardware stock visibility across the mapping fleet.
Maintained Jenkins + Ansible CI/CD pipeline for Python services; led ISO 9001 certification for safety-critical mapping software.
Managed technology partnerships with Intel, Nvidia, Audi, and Mercedes-Benz for sensor-to-cloud autonomous driving infrastructure.

Python Ecosystem Depth

Education

MBA, Business Administration — Quantic School of Business and Technology, 2017

PhD coursework, Physics (Quantum Chemistry, Solid-State Physics) — University of Central Florida, 2009–2013