Architect and technical lead for three production multi-agent AI platforms at Genentech,

serving pharmaceutical researchers across 57M+ records in regulated, high-availability environments.

• Designed a shared multi-client AI orchestration platform (LangGraph, FastAPI, MCP) with

Plan→Execute→Validate workflows, PostgreSQL-backed session memory, pgvector semantic search,

and config-driven client isolation, zero platform code changes per new application onboarded

• Built a natural language research interface integrating 20+ MCP servers and 150+ domain-specific

tools across 6+ siloed systems (PostgreSQL, MongoDB Atlas, Oracle), enabling access to 42M+

biological records and 170K+ sequences via a full-stack React 19 + TypeScript + Astro frontend

• Unified 6 in-vivo research systems across 4 database technologies behind a single natural language

interface with cross-system record linking and SSE streaming for real-time response rendering

• Reduced LLM token usage by 95% via cost-aware inference and two-stage summarization pipelines

while preserving 100% data coverage

• Achieved 10–100x data pipeline performance improvement by re-architecting REST integrations

to async direct database pipelines

• Defined cloud-native deployment standards, Kubernetes (AWS EKS), Helm, ArgoCD GitOps,

HashiCorp Vault, enabling zero-downtime releases across dev/tst/uat/prd environments

• Founded and scaled a global engineering team (SSF + Poland); set ADRs, technical standards, and influenced org-wide engineering practices for agent orchestration and AI platforms

Led cloud transformation strategy and enterprise architecture for Philips HealthSuite, moving

complex regulated clinical and imaging systems from on-premise to cloud-native PaaS.

• Strategized and architected HealthSuite Digital Platform (HSDP) cloud-native transformation,

designed PaaS services enabling interoperability, analytics, and AI integration across global

healthcare product lines

• Defined enterprise architecture frameworks and reusable reference models adopted across multiple

global engineering organizations for scalability, security, and regulatory compliance

• Architected federated data platforms supporting ML pipelines, advanced analytics, and clinical

decision support systems across distributed environments

• Implemented Multi-Factor Authentication (MFA) using TOTP and integrated with Philips IAM and

third-party auth apps (Google Authenticator, Microsoft, Authy)

• Designed and built cloud-native Vendor Neutral Archive (VNA) with AWS S3/Glacier lifecycle

policies for cost-effective long-term DICOM-PACS medical image archiving

• Built IAM microservices (Java, Spring Boot, PostgreSQL, LDAP) supporting authentication,

authorization, RBAC, OAuth, deployed on AWS EC2 with fully automated CloudFormation

• Championed cloud-native IAM standards, delivered secure auth/authorization across multi-tenant

regulated healthcare applications integrated with Validic, Salesforce, FitBit, Apple Watch ecosystems

• Conducted organization-wide AWS training for developers, accelerating cloud-native adoption

• Led architecture governance, design reviews, and technology selection across multiple orgs

• Mentored engineers; partnered with portfolio and product management to define digital roadmaps

• Defined roadmap and architecture for Philips Clinical Platform (HSDP Foundation) : IAM,

identity resolution, permission engine managing auth/authorization/consent across all Philips products

• Led Identity and Access Management (IAM) team of 6 engineers, single point of contact for

all IAM integrations across Philips Business Units (VHR, BDA, and others)

• Won Philips Worldwide Hackathon 2015 : built a working prototype from scratch in 24 hours

that was subsequently productized

• Developed FHIR-compliant Clinical Data Repository microservices (Java, Spring Boot, REST)

delivered as Docker containers on AWS/Cloud Foundry

• Played Product Owner for DICOM Print subsystem, customer discussions, roadmap, sprint demos

• Led a team of 10 engineers across multiple subsystems in Philips Clinical Platform

• Built Back End X-ray (BeX) platform modules: Viewing, User Guidance Service (UGS), Touch

Screen Module, on which all Philips X-Ray modalities built their applications

• Designed and optimized DICOM, HL7, and IHE interoperability subsystems ensuring regulatory

compliance (FDA, HIPAA) and clinical performance at global scale

• Led requirements engineering, architectural design (UIS/URS/UDS), and full-cycle delivery

of mission-critical imaging subsystems

• Performed performance tuning on image loading, scrolling, and algorithm optimization for

first-image latency in diagnostic workflows

• Designed, implemented, and performance-tuned SIENET Cosmos PACS radiology viewing

functionalities (Blow-Up, Magic Glass zoom, 2D interactive annotation tools)

• Optimized image loading and scrolling pipelines for large diagnostic dataset performance

• Built MSI installation packages for PACS viewer deployment

Stack: C++, VC++, MFC, Design Patterns, Rational Rose, Rational Purify