Lajos Moczar
Senior Agentic Engineer / Architect
Profile
30 years delivering production systems at scale, series A startups to Fortune 500 clients: payments and banking compliance (VocaLink/Mastercard, Custodia), telco and enterprise platforms (Swisscom, HERE Technologies), media and consumer (Blinkbox, MyTheresa, Starz Entertainment). Regulated & secure environment delivery experience: SEPA and BACS payment infrastructure, ISO 27001, SOC2, PCI-compliance. Solid experience with global, high-volume systems (HERE, Swisscom, Starz).
Three years building with LLM agents in production: prompt engineering, autonomous and semi-autonomous build harnesses, RAG pipelines, MCP servers and custom agent skills. One of my specialties is understanding why systems fail, and how to prevent that. Whether the builder is human or agent, the causes are the same in both cases: lightweight or sloppy specifications, abstract architectures, unhandled failure modes.
Unique ability to model complex systems across multiple dimensions: technology options, architecture, development velocity, testing, DevOps, management, security, reliability, failover, customer needs and management goals.
Agentic Engineering
- Streaming platform (client under NDA): Multi-tenant service v1 (billing, analytics, reporting, catalogue ingest, play tracking, PWA) delivered via autonomous agent pipeline using custom harness and skills, cross-agent review and automated PR handling. Outcome: Plan, design, build, test, review, bugfix for 6 core services, 10 kLOC, delivery in 4 days
- Publishing pipeline (client under NDA): Agentic pipeline for automated translation, verification and InDesign artifact generation for Amazon ebook publication. Pipeline capable of generating up to 20 books per week.
- DeFi notary service: Agent-built smart contract and API backend, end to end.
- RAG in production: Publishing catalogue data and live observability telemetry - embedding and chunking design, vector storage (LanceDB), retrieval tuning, log ingestion, evaluation harnesses, Python/FastAPI.
- Harness and agent design: Multi-agent decomposition, delegation boundaries, MCP servers, guardrails and validation schemas, agent failure and drift patterns, LLM routing and model selection, token cost and latency profiling.
- Outcomes: ~95% reduction in time vs. human-built; custom harnesses & skills + dark factory principles = ~30% reduction in agent-introduced coding errors.
Work Experience
Senior AI Engineer
- Agentic builds: semi- and fully-autonomous builds; multi-server streaming service, DeFi notary service, mobile apps, ebook generation pipeline; using automated PR processes, cross agent reviews, custom skills (plan, design, build, test, review, post-mortem, doc)
- RAG pipelines: publishing company book data & observability data; using embedding generation, vector storage, retrieval logic for assistant responses in live telemetry, log data ingestion.
- Technologies: Claude Code, Codex, Cowork, NodeJS, Typescript, AWS, Azure, Rust, Python, FastAPI, C++, Java, NextJS, React, Llama, Mistral, PyTorch, Gitea, InfluxDB, LanceDB
Senior AI & DevOps Engineer
- Using Claude Code & prompt engineering to develop one-click DevOps automations for multi-cloud, multi-region workload portability.
- Outcomes: 90% reduction in delivery time versus hand-coding.
- Audit, review, debugging, repair of client vibe-based SaaS systems.
- Technologies: Claude, ChatGPT, AWS, Azure, Python, Terraform, Ansible, GitHub, GitLab, Vercel, Netlify, Supabase, Kubernetes, Redis
Senior DevOps Engineer & Team Lead
- Designed and implemented global Azure architecture for a FinTech banking-compliance SaaS platform
- Implemented greenfield DevOps automation and tooling
- Developed testing tools for telephony/SIP infrastructure
- Provided technical strategy to leadership; trained and mentored development and operations teams and management of DevOps team
- Named technical contact for the platform’s largest banking client; brought into client meetings at their request to solve technical issues.
- Prompt-driven development of DevOps-related IaC tooling
- Technologies: ChatGPT, Copilot, Azure, AWS, Kafka, Mongo, ES, Python, Databricks, Azure DevOps, Terraform, Ansible, Asterisk
- Outcomes: fast, repeatable deploys with built-in release criteria; ISO 27001/SOC 2 compliant; MTTD < 10 min, MTTR < 60 min, MTTF > 6 months
Senior DevOps Engineer
- Implemented rigorous DevOps practices, increased automation, revamped GitHub pipelines to cut build times from 3-4 hours to < 1 hour, Packer for custom image builds, CI/CD pipeline to PXE-boot bare metals with custom images for cluster integration testing.
Senior DevOps / SRE
- Managed a Kubernetes-based e-commerce platform moving to AWS EKS
Senior Cloud Architect & Team Lead
- Led design and implementation of a global, multi-cloud self-service data platform for customers and internal teams, based on a customized Kubernetes.
- Enabled productized deployment model and efficient CI/CD processes for one-click deployments to a cloud and/or region of choice
- Developed tooling to improve financial visibility and support cloud usage billing to users.
- Technologies: AWS, Azure, OpenShift, Kubernetes, Jenkins, Helm, Kafka, Redis
- Outcomes: multi-cloud deployment design with fast deploy times; full Azure IaC; tooling for governance; cost usage for clients
Senior Systems Engineer
- Managed architecture & operations of a large global AWS estate (EC2, EMR/Spark, Databricks, Route53, Lambda, CF, Splunk, Redis) improving stability, observability, dashboards, alerting & achieving lower MTTD/MTTR.
Senior Enterprise Architect / Management Consultant
- Delivered architecture for IaaS/PaaS services (NSX, vSphere/vCenter, PXE, ScaleIO, NetApp), redesigned the CI/CD pipeline approach (Jenkins), and led real-time monitoring framework design with cross-cloud visibility
- Led feasibility studios on OpenShift, AWS, Azure and OpenStack; performed risk assessments on technology fit within functional project areas
- Reorganized the Swisscom architecture team to reduce churn with development teams
- Introduced a mid-level design specification practice that became a Swisscom standard
- Advised senior management on product roadmap and market positioning
Lead Architect & Backend Developer
- Designed and built a Java-based backend platform for a music-related social media startup: a multi-layered engine with Big Data & ML technologies for programmatic detection of various social and music trends. Technologies included Spring, AWS, Kafka, Storm, Cassandra, SolrCloud, Lucene, Hbase, Hadoop and YARN.
Senior Developer & Architect
- Revamped and improved a Big Data warehouse based on Hadoop and Hive for online radio station user and music data intelligence
- Designed a new corporate analytics data model and a more strategic approach to analytics within the context of business KPIs and goals
- Led development of a MapReduce, HBase and Storm-based approach for handling large volumes of user event information
CTO
- In charge of partner-facing engagements and technical solution development for customers
Technologies: Intershop, Hadoop, MapReduce, HDFS, Hive, Hama
CTO
- Product management oversight, development team leadership and technical alignment between customers, business and internal teams; hands-on backend development to fill knowledge gaps or deliver faster
- Hired and managed teams across two locations in Ireland
- Re-engineered a SaaS platform previously built by an offsite team to better meet business needs; worked with third-party providers to design and deploy a PCI-compliant, fully redundant hosting topology
Senior Solutions Architect
- Enterprise architect, team lead for SEPA (credit, debit, clearing) systems based on J2EE, WLS, SCA4J + solutions architect for related products
- Served as a technical pre-sales consultant for bank clients, resolving communications breakdowns between VocaLink and its clients, and architecting solutions to the technical issues they raised.
- Performed a technical evaluation for a potential acquisition by VocaLink
Owner, Consulting CTO
- Provided CTO services to two e-commerce startups who needed help in the initial stages of defining the technology portfolio, architecting products, engaging providers, hiring staff and developing IT plans and budgets
- For another e-commerce client: system to analyse eCommerce NLP search queries and evaluate relationship between results and purchase behaviour using Apache Mahout, Lucene, Solr
- Developed a multi-tenant SaaS eCommerce solution for a hosting provider
Enterprise Architect, Senior Developer, Team Lead
- Senior Java developer for BACS and SEPA services.
- Technologies: Mule, Hibernate, Spring, Maven, WebLogic, Oracle, SafeStore
Senior Java Developer, Enterprise Solutions Architect
- Worked on the next version of MuleSource's ESB product as an architect and developer, implementing registry capabilities, failover, and clustering features, as well as developing an Eclipse-based GUI IDE.
- Sent on-site to VocaLink in London as an enterprise architect/senior developer to help the client improve its implementation of Mule for its BACS system
Enterprise integration architect, Senior developer/team lead
- Implemented the Mule (now MuleSoft) ESB for a new movie download/viewing service (www.vongo.com). The system retrieved transactional data associated with financial, DRM and download activities, and funneled them into a data warehouse. The data was retrieved from both intranet and extranet sources in a variety of formats and connectivity mechanisms, and was designed to handle millions of transactions per day.
PUBLICATIONS
- Enterprise Lucene and Solr, Addison-Wesley Professional (2017) — enterprise search, text classification and ML, and productionalising retrieval systems
- Tomcat 5 Unleashed, Sams Publishing (2004)
- Cocoon Developer's Handbook, Sams Publishing (2002)
- 'Why Agile Isn’t Working: Bringing Common Sense to Agile Principles', CIO.com (2013)
- Named inventor on two Oracle patents
TECHNICAL SKILLS
- AI/ML: Claude Code, Codex, Cowork, MCP, LangChain, LanceDB, Cohere, Voyage, OpenAI, Llama, Mistral, PyTorch, RAG, embeddings, vector search, evaluation harnesses, prompt engineering, agent orchestration, guardrails, LLM routing, token cost optimisation
- Languages: Python, TypeScript, Java, C/C++, Rust, JavaScript, SQL
- Frameworks: FastAPI, React, Next.js, Node.js, Spring, Hibernate
- Data: Kafka, Spark, Elasticsearch, Solr, Lucene, MongoDB, Redis, InfluxDB, Hadoop, Databricks
- Infrastructure: Kubernetes, Docker, Terraform, Ansible, Helm, CI/CD, GitHub/GitLab, Jenkins
- Cloud: AWS, Azure, GCP, OpenStack, OpenShift
- Observability: Datadog, Prometheus, Grafana, ELK, Splunk
- Practices: ISO 27001, SOC 2, PCI-DSS, SRE, IaC, incident response, post-mortem