Paul Otto

Amplify Federal

• Provide platform engineering support for Second Front and Forterra.
• Build and maintain cloud infrastructure, Kubernetes platforms and deployment pipelines for their engineering teams.
• Improve observability, reliability and security across their environments.

Otto Ops LLC, Aurora, Colorado

• Built an end-to-end LLM fine-tuning pipeline with teacher-student trajectory generation, SFT, DPO, continued pretraining, SLERP merging, GGUF export and automated benchmark evaluation.
• Built a 4-bit QLoRA training harness for Gemma 4's fused-expert mixture-of-experts layers that fine-tunes a 19B-parameter model on one 32 GB consumer GPU.
• Developed a recursive language model (RLM and SRLM) training and inference framework with multi-trajectory generation, uncertainty-guided selection, context-length routing and map-reduce orchestration.
• Run self-hosted inference on llama.cpp in router mode with Flash Attention and a quantized KV cache. Prometheus, Grafana, Loki and Alertmanager watch it, and VRAM watchdogs catch spills to system memory during overnight training runs.
• From 2014 to 2017, consulted on big data pipelines and DevOps for enterprise clients. I built and ran big data infrastructure across three on-premise data centers, set up DevOps process and tooling for an advanced data analytics organization, and wrote and taught a Puppet curriculum on site for a Fortune 50 customer.

ClearBridge Revenue Integrity Partners, Denver, Colorado

• Build the cloud infrastructure, data architecture and automation for the company, including the ClearBridge Academy learning platform, built for HIPAA compliance.
• Architected a retrieval knowledge base for behavioral health billing regulations on LightRAG, with Neo4j knowledge graphs, Milvus vector search, bge-m3 embeddings and entity extraction across 18 domain entity types.
• Built the Academy's AI features, including Gemini TTS narration, Moshi speech recognition and GPU document processing with Docling and MinerU.

Clarity, Columbia, Maryland (remote)

• Led a distributed team of platform engineers on military systems.
• Lead engineer for the ADCP SRE organization at Kessel Run, US Air Force. Mentored military members, government civilians and contractors.
• Designed and built observability platforms on Kubernetes with Elastic, deployed with ArgoCD, FluxV2 and Terraform.

Raft, McLean, Virginia (remote)

• Wrote and presented proposals to government evaluation boards that won two SBIRs and the Cloud CITI 2.0 contract, with both option periods.
• Principal Architect and technical lead for Public Cloud ADCP at Kessel Run, US Air Force. Built Kessel Run's Zero Trust Network Architecture.
• Served as the subject matter expert on PlatformOne deployment patterns for a small team that launched an open source bare-metal private cloud edge node at the Air Force Shadow Operations Center, a test case for globally distributed C2.
• Led a growing team of software and systems engineers. Integrated Confluent Platform (Kafka), Istio and Kubernetes, deployed with Ansible, ArgoCD, FluxV2 and Terraform.

Ticketmaster, Hollywood, California (remote)

• Led engineering of the secrets management service in Q4 2019 that put Ticketmaster in CCPA compliance for January 2020.
• Managed a high-priority AWS cost management project and coordinated its rollout with more than 20 teams worldwide.
• Architected log usage metrics and chargeback services, and wrote the log usage tracking, metrics transmission and throttling code.
• Moved development teams from on-premise to EC2, from EC2 to Tectonic Kubernetes, and from Kubernetes to AWS EKS. Field-tested EKS and rolled it out as Ticketmaster's primary platform.
• Wrote backend services in Go and Python and mentored colleagues across the company in Go and cloud native practice.

TrackVia, Denver, Colorado

• Architected data services on Galera (Percona XtraDB Cluster), MySQL 5.6, HAProxy with Keepalived, and Hadoop.
• Architected the AWS infrastructure for TrackVia's new product, including VPC, EC2, CloudFormation, IAM, RDS, EMR and S3, automated with Boto, Ansible and Puppet.
• Managed infrastructure, data services and production operations, and mentored engineers in operations and DevOps practice.

GoDaddy, Denver, Colorado

• Supported CentOS Linux and MySQL on more than 2,400 high-traffic servers, and automated operations with Bash, Perl, PHP and Python.
• Planned upgrades across MySQL 4.0, 4.1, 5.0 and 5.1, and from MyISAM to InnoDB.
• Led a team of MySQL engineers building an internal service to manage every MySQL asset, and set the department's git practices.