James Westover

• LinkedIn: @jameswwestover
• GitHub: @westover
• Scite: Author Page
• Portfolio: westover.dev
• Westover Labs: westoverlabs.de
• Consulting: westover.expert
• Developer Tools: breadfactory.app
• Email: james@westover.dev
• Phone: +1-510-207-8194

Summary

I'm a seasoned systems architect and Python expert leveraging AI as a force multiplier to accelerate development and innovation. With deep experience in mapping, autonomous driving, and distributed systems, I now lead virtual development teams powered by Claude automation — orchestrating AI agents to rapidly build, deploy, and scale production systems. From platform modernization at Uber to AI-driven development at Westover Labs, I specialize in cost-effective, privacy-first infrastructure that empowers developers to build at home without metered cloud constraints. My recent work includes launching subscription iOS apps with open-source ML models, FastAPI ONNX inference services, and home-hosted infrastructure using Cloudflare, GitHub, Tailscale, and Ansible.
Specialties: Python, AI-Driven Development, Distributed Systems, FastAPI, React Native/Expo, ONNX Inference, Home Hosting Infrastructure, Privacy-First Architecture, Developer Tools, Cloud Architecture (AWS), Mapping, GIS, ML Infrastructure.

Technical Qualifications

• Languages & Scripting: Deep expertise in Python (2.4–3.x) as a primary backend language. Production experience with JavaScript (ES6), TypeScript (infra/tooling), SQL, Golang, Java, C++, Shell scripting (Bash, Zsh, Windows CMD), FORTRAN (77, 90, 95).
• Backend & Python Libraries: Django, Flask, Tornado, FastAPI, WSGI, Celery, SQLAlchemy, Alembic, Pydantic, Marshmallow, Requests, Pytest, Tox, PDM, Virtualenv, AsyncIO, Multiprocessing, Threading, boto3, Click, Typing, Dataclasses, logging.
• ML & AI Frameworks: PyTorch, TensorFlow, Scikit-learn, NumPy, SciPy, Pandas, Matplotlib, OpenCV, ONNX Runtime. Extensive experience with Claude AI for development automation and AI-augmented workflows.
• Mobile Development: React Native, Expo, TypeScript, ARKit, CoreLocation, StoreKit 2 (subscription management), on-device ML inference, cross-platform mobile architecture.
• Serialization, RPC & Messaging: gRPC, Apache Thrift, Protobuf, Apache Avro, Apache Parquet, Kafka, RabbitMQ, TChannel (primary contributor to Python 2→3 migration).
• Streaming & Big Data: Apache Kafka (including Confluent Platform), Apache Storm, Apache Spark, Apache Flink, Apache Airflow, Hadoop, MapReduce, YARN, Hive, HBase, Druid, Zookeeper, Presto, ClickHouse, Amazon Kinesis, EMR, Athena, S3 Select, Google Bigtable; also built internal workflow and data orchestration systems (e.g., Warehouser at Sonobi, Weatherman at HERE).
• Cloud & Infrastructure: AWS (ECS, EC2, S3, SSM, SageMaker, Lambda, CloudFormation, CloudWatch, IAM), GCP (Bigtable, Compute Engine), Azure (VMs, Storage), on-prem and hybrid infrastructure (including SaltStack, Mesosphere/DCOS, internal Uber orchestration systems, and private cloud deployments using XenServer and Eucalyptus); familiar with OpenStack. Expert in privacy-first home hosting infrastructure using Cloudflare (edge networking, tunnels, DNS), Tailscale (mesh VPN, zero-trust networking), and Ansible (infrastructure as code). Used secrets management tools including AWS KMS, GitHub Secrets, and GCP Secret Manager. Docker, Kubernetes, Envoy, Terraform, Ansible, Jenkins, GitHub Actions, GitLab CI, Nginx, Varnish, HAProxy, Prometheus, Grafana, Datadog, ELK Stack, CDN integration, Load Balancing, Auto Scaling, Immutable Infrastructure.
• Data & Storage Technologies: PostgreSQL, MySQL, Oracle, Microsoft SQL Server, Cassandra (including multi-region deployments, tuning tombstone behavior, and managing consistency models such as QUORUM, LOCAL_QUORUM, and ONE for read/write performance and availability tradeoffs), Redis, Memcached, HBase, Hive, Google Bigtable, Amazon Redshift, Elasticsearch, Apache Solr, Druid, S3, Zookeeper; experienced in schema design, indexing strategies, and distributed storage tradeoffs at scale.
• Geospatial Technologies: PostGIS, OpenLayers, GeoJSON, GeoServer, GDAL, Shapely, GEOS, Proj, Leaflet, Mapbox, QGIS; experience with Uber’s internal mapping stack including S2 and H3 cell indexing, real-time geohashing, sensor fusion pipelines, and off-route detection. Built and optimized tile services (including vector tiles) and on-device map matching infrastructure for mobile inference. Contributed to routing optimizations including custom solutions to variations of the Traveling Salesman Problem (TSP) for fleet and logistics use cases.
• Autonomous Systems & Sensor Infrastructure: Deep experience in autonomous driving and sensor fusion, with a focus on end-to-end systems bridging real-time vehicle data and distributed backend platforms. Worked extensively with ROS (Robot Operating System) in mapping vehicles; contributed to pipeline design for edge data capture, onboard preprocessing, wireless uplink, and cloud-based ingestion and analysis. Led technical efforts across HERE, Intel, Nvidia, Mercedes-Benz, and Audi, specializing in sensor-to-cloud communication and fleet-scale data infrastructure. Hands-on experience with HD map generation, fleet telemetry pipelines, and real-time localization systems.
• HPC & Scientific Computing: Extensive experience in high-performance computing for scientific applications, including Monte Carlo simulations, molecular dynamics (LAMMPS), and electronic structure calculations (VASP). Built and maintained Beowulf-style compute clusters using ROCKS, including diskless node deployment and MPI-based parallelization with MPICH2 and OpenMPI. Proficient with OpenMP, multithreading, and custom HPC toolchains. Designed and implemented real-time microsecond-precision timing systems for laser cooling in Bose–Einstein condensate experiments; strong background in signal synchronization, hardware interfacing, and experimental system design. Familiar with analytical and modeling tools including MATLAB, Mathematica, Maple, LabView, and SPSS.
• Engineering Process & Quality Systems: Trained in Agile, Scrum, and Lean Six Sigma methodologies; experienced in applying continuous improvement and iterative development practices to engineering orgs. Heavily participated in ISO 9001 certification efforts for a prior organization, implementing quality management systems and process documentation standards to support scalable, auditable software delivery.
• Leadership, Mentorship & Communication: Experienced engineering manager and technical leader with a strong background in team building, mentorship, and cross-functional execution. Currently pioneering AI-augmented development teams at Westover Labs, serving as Head of Engineering to orchestrate Claude AI agents as virtual team members — demonstrating how individual developers can achieve team-level productivity through strategic AI orchestration. At HERE, grew the Capture Systems Software team from 2 to 15 engineers, structuring it into multiple subteams (frontend, mobile/tablet, backend, DevOps, and embedded) to support a wide range of autonomous driving projects. Served as architect and IC across backend, DevOps, and embedded systems while leading people and delivery. At Uber, built the Location Platform team from the ground up, making key hires who continue to thrive in senior roles. Played key roles in cross-org collaboration and alignment efforts at both companies, coordinating closely with platform, data, and product teams to deliver complex, multi-stakeholder infrastructure. Former undergraduate instructor and curriculum designer with experience developing and teaching lab-driven STEM courses. Skilled in aligning technical vision, mentoring both human and AI team members, and scaling team culture and process.

Projects

• Westover Labs

  • Claude Automation & AI-Driven Development (nds_server)

    • Built virtual development team powered by Claude AI agents, acting as Head of Engineering to orchestrate autonomous bot developers.
    • Achieved rapid development cycles by leveraging AI as a force multiplier — automating code generation, testing, deployment, and infrastructure management.
    • Architected AI-driven workflows for full-stack development including React Native, FastAPI, and ONNX inference services.
    • Demonstrated how individual developers can achieve team-level output through strategic AI orchestration and tooling.
    • Pioneered privacy-first, cost-effective home hosting infrastructure using Cloudflare, GitHub, Tailscale, and Ansible.

    Technologies: Claude AI, Python, FastAPI, React Native, Expo, ONNX Runtime, Docker, Ansible, Tailscale, Cloudflare, GitHub Actions

    Skills: AI Orchestration, Team Leadership, DevOps Automation, Infrastructure as Code, Rapid Prototyping, AI-Augmented Development

  • Ghost Hunter iOS Subscription App

    • Rebuilt and launched subscription-based AR ghost hunting game on Apple App Store.
    • Integrated open-source ML models for real-time object detection and inference using ONNX Runtime.
    • Deployed cost-effective FastAPI inference services with home-hosted infrastructure managed by Ansible and Tailscale.
    • Achieved production-grade reliability while avoiding metered cloud costs through strategic use of Cloudflare edge networking and GitHub-based CI/CD.
    • Implemented subscription management using StoreKit 2 and server-side receipt validation.
    • Developed using React Native and Expo for rapid iteration and cross-platform capability.

    Technologies: TypeScript, React Native, Expo, ARKit, ONNX Runtime, FastAPI, Cloudflare, Tailscale, Ansible, GitHub Actions, StoreKit 2, Open Source ML Models

    Skills: Mobile Development, AR, On-Device ML, Cost Optimization, Subscription Monetization, Home Infrastructure, Privacy-First Architecture

  • Breadfactory.app - Home Developer Tools Platform

    • Building platform and tooling to empower at-home developers to take ideas to production without cloud vendor lock-in.
    • Emphasis on privacy-first development, cost transparency, and freedom from metered cloud services.
    • Providing patterns and tools for home hosting with professional-grade infrastructure using Cloudflare, Tailscale, and open-source technologies.
    • Enabling developers to maintain control over their data and infrastructure costs while achieving scalable, production-ready deployments.

    Technologies: Python, FastAPI, Docker, Ansible, Tailscale, Cloudflare, Open Source Tools

    Skills: Developer Tools, Platform Engineering, Infrastructure Design, Privacy Engineering, Cost Optimization, Developer Education

  • Transit Authority Train Project (Bid Not Selected)

    • Designed a solution to monitor passenger and driver behavior using RTSP camera streams.
    • Built a mock-up dashboard for behavior visualization.
    • Architected a scalable backend with AWS Rekognition and SageMaker for real-time image analysis.
    • Integrated with Amazon Keyspaces and S3 for metadata and image storage.
    Technologies:Python, Django, Docker, RTSP, AWS Rekognition, AWS SageMaker, Amazon Keyspaces, S3, Elastic Container Service
    Skills: Proposal development, system architecture, computer vision, cloud infrastructure
• Uber

  • Navigation

    • Contributed to one of Uber’s largest backend services (approx. 10,000 nodes per data center), powering all routing and mapping across Uber apps.
    • Introduced support for off-route cross-street links to the route object, reducing p95 off-route detection time from ~5 minutes to ~2min
    • Enhanced navigation engine heuristics for off-route links, enabling faster rerouting triggers.
    • Developed Python-based simulation tools to test route deviation detection at scale, boosting simulation capacity from 500 to 10,000 trips per run.
    • Maintained J2ObjC transpilation system to share Java routing logic with iOS app.
    • Diagnosed and mitigated critical GPS clock offset issue affecting many devices.

    Technologies: Java, J2ObjC, Python, Android, iOS, Multiprocessing, GPS, Sensor Fusion, Map Matching, Routing Engines

    Skills: Troubleshooting, Simulation, Cross-Platform Engineering, Incident Response, Backend Scalability, Cross-team collaboration, Cross-organizational collaboration

  • Location Platform

    • Led Location Platform team, increasing global GPS+sensor data ingestion success from ~80% to 97%+.
    • Built GDPR-compliant location collection, enabling rider-driver sharing
    • Improved data reliability for crash detection and “RideCheck” safety (RideCheck blog).
    • Launched ShadowMaps project for multipath-aware geospatial correction (ShadowMaps blog).
    • Oversaw anti-location-fraud efforts, saving $30M+ via spoofing detection.
    • Scaled Cassandra infrastructure to support global ingestion and high QPS (2–10M QPS).

    Technologies: Python, Cassandra, Kafka, Thrift, H3, S2, Protobuf, Distributed Systems, Real-Time Location, Privacy Compliance, Geospatial Analytics

    Skills: Team Leadership, Compliance, Data Quality, High-Scale Systems, Anti-Fraud, Cross-team collaboration, Cross-organizational collaboration

  • Python Platform

    • Founding engineer for Python Platform; supported Python development company-wide.
    • Provided mentorship and office hours for debugging, tuning, and compatibility issues.
    • Led Python 2→3 migration; mapped and updated dependencies impacting 90%+ of production services.
    • Major contributor to tchannel-python (Tornado-based Thrift RPC transport).

    Technologies: Python 2.7–3.x, Tornado, TChannel, Thrift, Dependency Management, Refactoring, Distributed Systems, Open Source Collaboration

    Skills: Mentorship, Migration Leadership, Open Source, Codebase Refactoring, Developer Enablement, Cross-team collaboration, Cross-organizational collaboration

  • Postmates Integration

    • Maintained legacy Postmates systems during integration, ensuring service continuity and extended on-call support.
    • Integrated Postmates customer identity data for compliant data deletion for unlinked accounts.
    • Worked with Go and Uber’s internal schemaless datastore for identity and compliance data flows.

    Technologies: Go, Schemaless, Identity Systems, Data Deletion Compliance, System Integration, Legacy System Maintenance

    Skills: Integration, Reliability, Compliance, Legacy Support, Cross-Team Coordination, Cross-team collaboration, Cross-organizational collaboration

• Postmates

  • Python 2 to 3 Migration at Scale in Production

    • Led migration for Postmates’ monorepo, ensuring Python 3 compatibility across a large, interdependent codebase.
    • Orchestrated distributed test jobs across Kubernetes pods, reducing CI build times from 40–50 to 10–15 minutes.
    • Collaborated with infra and observability teams to improve deployment quality, integrating Envoy and OpenTelemetry.
    • Debugged and refined production microservices in high-traffic, latency-sensitive environments.
    • Participated in on-call rotations and postmortems, driving reliability improvements.
    • Lived through the full AWS → GCP cloud migration. All Kubernetes work was CLI and YAML-first — kubectl, k9s, deployment manifests. Never touched the console. StatefulSets for databases inside k8s were the genuinely hard part.

    Technologies: Python 2.7–3.x, Kubernetes (kubectl, k9s), Docker, Envoy, OpenTelemetry, Prometheus, Grafana, GitLab CI, Monorepo Architecture, Distributed Systems, PostgreSQL, pgpool, PgBouncer, Django, GCP

    Skills: Migration Leadership, CI/CD Optimization, Cross-Team Collaboration, Reliability Engineering, Observability

• HERE Technologies

  • Leadership and Team Building

    • Built and scaled engineering team from 2 to 15, structured by backend, frontend, embedded, mobile, and DevOps.
    • Led across in-car platforms, API design, infrastructure, and visualization.
    • Provided architectural and technical mentorship while acting as people manager.
    • Supported ISO 9001 certification, including QMS implementation and documentation.
    • Facilitated Agile practices (Scrum of Scrums, release trains, cross-team planning).

    Technologies: Python, Jira, Team Processes

    Skills: Agile, Lean Six Sigma, ISO Certification, Technical Leadership, Organizational Scaling

  • Fleetvision 2

    • Developed Tornado-based backend APIs for in-vehicle stack integration.
    • Built React-based frontend for driver/engineer tablet interface.
    • Integrated real-time GPS, IMU, and camera telemetry for fleet tracking.
    • Built OTA update and diagnostics pipeline with Jenkins and Ansible.
    • Map visualizations inspired custom web app we developed for video wall (16 tv array) in Berlin office

    Technologies: Python, React, Tornado, PostgreSQL, SQLAlchemy, PostGIS, Redis, Elasticsearch, Openlayers, Geoserver

    Skills: Real-Time Data, Stakeholder Management

  • Fleetvision Data (FVD)

    • Designed/implemented lambda architecture for scalable vehicle data ingestion and analytics.
    • Used Apache Storm for streaming, Kafka for ingestion, Elasticsearch/PostGIS for storage/search.
    • Ran the largest single unsharded PostgreSQL/PostGIS instance in production — the entire HERE world map in one DB. The real challenge: spatial queries at APAC/EMEA/LATAM/NORAM partition boundaries for autonomous vehicles driving across shard edges in real time. Pre-HDFS-GIS era; invented the approach before tooling existed.
    • Apache Thrift as universal IDL: one schema generated Python services, C++ ROS nodes, and JS types. Storm topology classified tuples via Thrift-typed events — typed contract end to end across the stack.
    • GeoServer for tile generation: raster source → vector tiles → custom Thrift objects consumed by the fleet visualization stack.
    • Built “Weatherman” orchestration for dynamic map reprocessing and QA.
    • Maintained Jenkins+Ansible CI/CD pipeline for seamless deployment/testing.
    • Supported Fleetvision 2, decoration display webapp, Spotcam Monitor, HT-Route Creator downstream.

    Technologies: Python, Apache Storm, Kafka, Elasticsearch, PostgreSQL, PostGIS, Apache Thrift, GeoServer, Jenkins, Ansible

    Skills: Data Pipeline Design, Orchestration, CI/CD, Scalable Analytics

  • ADAS Prototyping

    • Developed PoC pipelines for ADAS using real-time sensor data.
    • Integrated Python processing with ROS for rapid prototyping of perception/localization.

    Technologies: Python, ROS

    Skills: Rapid Prototyping, Sensor Fusion, Perception, PoC Development

  • HT4: Next-Generation Mapping Rig

    • Integrated JPEG-compressed sensor streams from FPGA-based encoder boards.
    • Designed software abstraction for stream handling and metadata tagging.
    • Coordinated embedded/backend testing for reliable cloud ingestion.

    Technologies: Embedded Systems, FPGA, Python

    Skills: Hardware-Software Integration, Testing Coordination, Streaming Data

  • Corporate Collaborations

    • Worked with Intel, Nvidia, Audi, Mercedes-Benz for sensor-to-cloud ADAS/autonomous integrations.
    • Coordinated cross-company engineering for real-time data collection and map building.

    Technologies: Sensor Integrations, Real-Time Systems

    Skills: Cross-Company Collaboration, Real-Time Systems, Autonomous Driving

  • Inventory Management System

    • Designed/built QR code-based system for rapid inventory scanning and auditing.
    • Enabled real-time stock/component usage visibility across fleet.

    Technologies: Python, Flask

    Skills: Inventory Workflow, Process Automation, Systems Design

  • HT Route Creator

    • Custom webapp for route generation for mapping fleet
    • Legacy windows application contractors created provides routes for mapping fleet
    • Migrated physical windows machine into cloud instance to improve reliability and scaling
    • Created a consistent set of generated routes across all fleet managers
    • Replaced client desktop application with web application with pluggable apis to allow for replacement of routing engine

    Technologies: Python, React, Tornado, PostgreSQL, SQLAlchemy, PostGIS, Redis, Elasticsearch, Openlayers, Geoserver

    Skills: Software Lifecycle Management, Stakeholder Management

• Sonobi

  • Dynamix

    • Developed web scraping/automation framework to collect bid pricing and market data across platforms.
    • Used Python, BeautifulSoup, and Requests for HTML parsing/API interaction.
    • Reverse engineered and validated undocumented APIs using Wireshark and diagnostics.

    Technologies: Python, BeautifulSoup, Requests, Wireshark

    Skills: Web Automation, API Reverse Engineering, Data Collection, Diagnostics

  • Warehouser

    • Built data warehousing pipeline aggregating impression-level data across campaigns.
    • Aggregates log events per impression, improving reporting/optimization granularity.
    • Implemented ETL using S3, Redshift and Python, RabbitMQ, with robust logging/validation.

    Technologies: Python, Redshift, S3, RabbitMQ

    Skills: ETL, Data Modeling, Data Quality, Reporting

  • Probabilistic Group Overlap Estimation

    • Designed system to estimate overlap between large user groups with probabilistic data structures.
    • Combined HyperLogLog and MinHash sketches for efficient intersection approximation.
    • Used redis to store HLL objects and Minhash objects
    • Used Python and NumPy, with unit testing to ensure statistical confidence.

    Technologies: Python, HyperLogLog, MinHash, NumPy, Redis

    Skills: Probabilistic Algorithms, Statistical Validation, Performance Optimization

Experience

Education

Quantic School of Business and Technology

Master of Business Administration (MBA), Business Administration and Management, General
2017 - 2017

University of Central Florida

PhD, Physics
2009 - 2013

University of Central Florida

MS, Physics
2009 - 2010

University of Rochester

BS, Physics, Applied Mathematics
2006 - 2009

Publications

Metallization of the β-SiC(100) 3\times2 Surface: a DFT Investigation Link: Paper PlumX Metrics

Electronic Structure of an organic/metal interface: Pentacene/Cu(110) Link: Paper Scite Report

Initial velocity V-shapes of young asteroid families

Bryce T. Bolin, Kevin J. Walsh, Alessandro Morbidelli, Marco Delbo

• Acknowledgement: “BTB would like to acknowledge James W. Westover for thought-provoking discussions on the implementation of large-scale computing resources and algorithms that were used in the completion of this work.”