James Westover

Head of Tech & Data — Backend Architecture, Data Strategy, AI Enablement, Team Leadership

Profile

Technical leader who has built, scaled, and shipped at the exact intersection LANCH needs: food delivery infrastructure at Uber scale, data-driven product development, AI automation, and high-performance team leadership. At Uber I operated the navigation platform that powered every ride and delivery globally — 10M+ requests per second, 1400-node services per data center, 80+ node Cassandra clusters. I know exactly what breaks when a food delivery platform scales internationally, because I have been inside the machine that serves hundreds of millions of deliveries.

Now I build AI systems that do the work. At Westover Labs I run a 12-persona autonomous AI agent team that ships production code 24/7 — routing tasks by intent, filing PRs, running Jira workflows, and managing its own knowledge graph. This is not a demo. It is production infrastructure I use daily, built on FastAPI + PostgreSQL + React + TypeScript — the same stack LANCH runs.

50/50 hands-on and leadership. Physicist by training, production engineer by trade. MBA who speaks business. Staff engineer who has seen failure at scale and knows how to prevent it.

Why This Matters for LANCH:

I have scaled what LANCH is building. Uber's delivery platform (Uber Eats + Postmates) is the system LANCH is competing against. I built the location, routing, and data infrastructure that powered it. I know the failure modes — from database hotspots during dinner rushes to GPS spoofing from fraudulent drivers — before they happen.
Your exact tech stack is my daily driver. FastAPI, PostgreSQL, React, TypeScript, GitHub Actions, Docker — I write production code in this stack every day. Not "familiar with" — actively building.
AI enablement is not a future project for me. I already run an autonomous AI agent team in production. I can bring that capability to LANCH — automating internal workflows, data pipelines, and developer tooling — starting week one.
Data architecture at real scale. I have operated Cassandra clusters ingesting millions of GPS points per second, built lambda architectures processing 250M+ events per day, designed Kafka pipelines for real-time fleet analytics. BigQuery + dbt is a simplification of problems I have solved at much larger scale.
International expansion is familiar territory. Uber is a global company. I have worked on multi-region deployments, GDPR-compliant data architectures, and cross-timezone team coordination. Scaling LANCH across new markets is a playbook I have run before.
Berlin-based, available immediately. German national. No visa, no relocation needed.

Experience

Neher Data Systems / Westover Labs — Berlin, Germany

CTO, Neher Data Systems · January 2021 – Present

Founder, Westover Labs · February 2026 – Present

Built and lead a 12-persona autonomous AI agent team that ships production code continuously — agents handle task routing, PR creation, Jira management, knowledge graph updates, and code review without human intervention on routine tasks. This is the AI enablement capability I would bring to LANCH.
Own the full tech stack end-to-end: FastAPI backend services, PostgreSQL + pgvector data layer, React Native mobile apps, TypeScript frontends, Ansible infrastructure-as-code, GitHub Actions CI/CD, Cloudflare edge infrastructure.
Defined and executed product roadmap across three concurrent product lines (iOS apps, API services, internal platforms) — translating business requirements into technical delivery plans, shipping to App Store, managing third-party vendor integrations.
Built data-driven decision infrastructure: knowledge graph with Bayesian reasoning (PyMC), semantic search (pgvector/384-dim embeddings), real-time observability (OpenTelemetry), and automated anomaly detection.
Coach junior developers and AI agent contributors — run sprint planning, backlog grooming, code review, and retrospectives. Established engineering culture emphasizing speed, quality, and pragmatism.
Tech hiring: recruited and onboarded contractors; established code review standards and contribution protocols.

Uber — San Francisco, CA / Remote

Staff Software Engineer (Navigation, Maps) · August 2021 – October 2022

Owned on-device map matching and off-route detection for Uber's navigation platform — the company's largest deployment with tens of thousands of nodes per data center. Rolling deploys took a full week to propagate. Code runs on 30M+ driver devices serving every ride and delivery globally.
Off-route detection P95 was 5 minutes, P99 10 minutes — riders and drivers stuck on the wrong route for that long is a terrible experience. Led the research effort: hundreds of hours of simulation, built and optimized testing infrastructure to answer questions at massive scale. Full PRD/ERD process, built an org-level coalition to ship Phase 1.
Worked cross-team with Maps, my former Location Platform team, and engineers sourced through Uber's internal gigs system — no dedicated headcount, driven by initiative. Physics and math background directly applicable: audience sizing for experiments across billions of trips requires the same statistical thinking as simulation design.
Built Python-based simulation software — scaled capacity from 500 to 10,000 trips per run, enabling data-driven evaluation of route-matching algorithm changes before deploying to the fleet.

Staff Software Engineer (Eats / Core Identity) · February 2021 – August 2021

Solved the Postmates data deletion problem: how do you connect users who asked Uber to delete their data if they never consented for Uber to have it? Designed cross-platform identity resolution for unmigrated Postmates users — GDPR-compliant data deletion across multi-cloud distributed databases.

Senior Software Engineer (Python Platform, Office of CTO / UberAI) · Mid-2019 – May 2020

Founded the Python Platform team. Responded to an Office of CTO call-to-arms for Python 2→3 migration. Advocated internally for promoting Python to a supported language — all the data science teams used it, and the company was wasting enormous effort converting Python notebooks into Scala, Java, or Go.
Led migration across 400+ microservices: analyzed full dependency graph, targeted top 200 deps covering 90%+ of production services. This is what “high-level tech decisions across an entire company” looks like in practice.
Sole on-call for all of Python Platform — never got the promised additional staff (profitability push delayed hiring, then COVID). Mentorship and office hours for debugging and architecture decisions across the entire engineering org.

Software Engineering Manager (Sensing & Perception / Location Platform) · February 2018 – Mid-2019

Founded and built Location Platform team from scratch. Team started in Sensing & Perception (Maps), reorged into Office of CTO, then into UberAI. Made key hires whose careers I’m proud of — one grew from engineer to staff engineer leading teams over 8 years, another was leading a team when I returned in 2021. Conducted thousands of structured interviews across my Uber tenure (3–7 scorecard-level interviews per week, more as EM).
Cross-functional consulting across the entire location stack — fraud detection and device fingerprinting, terrestrial mapping fleet data, autonomous vehicle programs, beacon device hardware (plug-in signs planned with sensors), Mesos container management with Debian packaging for legacy OS support, Jenkins-based internal PyPI, Uber Listen and driver surveys. Location Platform was the connective tissue across dozens of teams.
Increased global GPS+sensor data ingestion success from ~80% to 97%+ — directly improving delivery ETA accuracy and driver-rider matching quality.
$30M+ annual savings from GPS anti-fraud program — detecting location spoofing that food delivery platforms (including LANCH's competitors) are increasingly vulnerable to as they scale.
Scaled Cassandra infrastructure to 80+ nodes per DC, 2–10M QPS insert rate globally — capacity planning, multi-region deployment, consistency model tuning. This is the data architecture muscle that makes international expansion work.
Built GDPR-compliant location pipeline (Project Highlander) — consolidated data into a single source of truth for privacy compliance. Participated in legal proceedings as technical expert.
Launched ShadowMaps: multipath GPS correction using building geometry, extended to Jump bikes for maintenance routing in urban canyons — because bikes disappear in the same city blocks where food deliveries happen.

Postmates Inc. — San Francisco, CA

Staff-Level Software Engineer (Core Infrastructure) · June 2020 – February 2021

Joined Postmates one week before the Uber acquisition was announced. A food delivery platform at a comparable growth stage to LANCH — fast-moving, venture-backed, 1M+ LOC Python monorepo. This is where I learned what a food delivery backend actually looks like under the hood.

Core infrastructure engineer for a food delivery platform. Staff-level IC on a 5-person team (including manager). Worked hand-in-hand with eng manager, with 40% and 20% allocations from two other engineers — small team, massive scope.
Completed Postmates’ Python 2→3 migration right as the Uber acquisition closed — coordinated across 100+ engineers, managed dependency sequencing across 3 coexisting Django versions while delivery traffic ran in production.
Kept the entire Postmates backend alive solo on-call 24/7 for 6 months after Uber sunsetted the platform. Every minute of downtime meant missed deliveries during the transition period.
Delivered 4x CI build time improvement (40–50 min → 10–15 min) by designing distributed test execution across Kubernetes pods. Built blue-green deployment infrastructure (Envoy) and OpenTelemetry observability.

HERE Technologies — Berlin, Germany

Software Engineering Manager (Capture Systems, Highly Autonomous Driving) · April 2016 – February 2018

Built team from 2 to 14 engineers — structured across backend, frontend, mobile, DevOps, and embedded subteams. Managed concurrent delivery across 8+ programs. Coordinated with Intel, Nvidia, Audi, and Mercedes-Benz on joint engineering.
Implemented ISO 9001 QMS — quality management for safety-critical software. Ran Agile/Scrum of Scrums across subteams.
Designed vehicle-to-cloud computer vision infrastructure: custom FPGA image compression, edge preprocessing, LTE uplink optimization. Cloud architect for autonomous driving data pipelines.

Senior Software Architect · April 2015 – April 2016

Built Fleetvision Data: lambda architecture processing 250M+ GPS events/day using Kafka + Apache Storm + Elasticsearch + PostGIS. This is real-time fleet analytics — the same class of problem food delivery platforms solve for driver tracking.
Built Tornado-based Python microservices, real-time fleet tracking with React frontend, and PostGIS-backed route optimization tools.

Sonobi — Orlando, FL

Software Engineer · December 2013 – January 2015

Built data warehousing pipeline processing 250M+ daily events (S3, Redshift, RabbitMQ). Designed probabilistic audience overlap estimation with HyperLogLog/MinHash — the same techniques used for food delivery demand forecasting.
Led AWS-to-bare-metal migration: reduced per-unit cost by 10x and cluster size to 1/4 — infrastructure cost optimization at scale.

Data & AI Capabilities

Education

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

PhD coursework, Physics (Quantum Chemistry, Condensed Matter) — University of Central Florida, 2009–2013. Published researcher.