Senior Full-Stack Engineer · Cloud Platform · Event-Driven Systems · AI Agent Platforms
First software engineer on a dental-manufacturing automation team, with 5 years building cloud platforms, real-time applications, and production AI systems. I set the architecture, infrastructure-as-code, authentication, and frontend standards supporting 80+ backend services and 15+ React applications — from an event-driven platform processing 1M+ events a day to an Amazon Bedrock agent platform that turned week-long investigations into self-serve answers.
My work sits where machines, cloud infrastructure, operations teams, and customer-facing systems meet: ingesting telemetry, designing backend services, automating environments, and shaping dashboards that make operational decisions faster.
The first software engineer on the Automation team, a hardware engineering group — I joined mechanical and controls engineers bringing connected devices online across five sites and built the cloud telemetry backend and the application platform above it. I own architecture and delivery end to end: design the systems, deploy the infrastructure, then guide two engineers through implementation and review changes before production.
Architected the topic-routed ingestion platform over AWS IoT Core and Kafka/MSK processing 1M+ events a day — fanning out to webhooks, live WebSocket clients, a time-series store, and 100+ DynamoDB tables, cutting operational visibility from minutes to seconds.
Established the Terraform + Lambda delivery standard — a uniform repo scaffold, GitLab CI, and multi-account promotion, now the default across 80+ backend services — plus the shared React 19 / MUI / Vite design system behind 15+ applications. New services and new apps both start from a working shell instead of a blank page.
A production Bedrock agent platform answering natural-language questions over live operational data, fleet analytics over 3,500+ machines with idle detection, and closed-loop ChatOps that turns an alert into a one-click remediation from Microsoft Teams.
I am strongest in product-facing platform work: building the services, interfaces, and automation that connect edge devices, cloud infrastructure, internal users, and business outcomes.
REST APIs, microservices, event contracts, cloud storage, data transformation, service orchestration.
A shared React 19 / MUI / Vite design system, token-driven light/dark theming, real-time WebSocket dashboards, operator workflows, chart-heavy reporting.
Standardized Terraform + Lambda patterns, multi-account promotion, GitLab CI/CD, authentication and RBAC, deployment hygiene, production support.
Amazon Bedrock, planner-routed agent orchestration and tool-use, NL-to-SQL over Athena, human-in-the-loop gating on writes, prompt engineering.
Where production details are confidential, these are framed by problem type, architecture, and capability. The two leading platforms — the IoT device-management and OTA control plane and the unified data-exchange backbone — I designed, built, and own as the sole engineer. The rest span a manufacturing execution system and workflow engine, a production Bedrock agent platform, the shared React design system, fleet analytics, ChatOps remediation, and the telemetry and authentication infrastructure beneath them, where I set the architecture and led delivery.
The cloud control plane for a fleet of CNC milling machines — provisioning devices, pushing firmware and configuration over the air, and reconciling what each machine actually runs against what the cloud intends. Designed, built, and owned end to end.
A topic-routed pub/sub backbone ingesting all factory MQTT traffic and fanning out to webhooks, live WebSocket clients, and a time-series store (Timestream for InfluxDB) — the ingestion layer the rest of the platform reads from.
An event-driven MES of 19 serverless microservices that digitized a paper-based production process — a configurable workflow engine where every step emits structured events, resurfaced as LLM-generated dashboards and real-time KPIs for plant leadership. I set the architecture and infrastructure and built the core services, with the team delivering alongside me.
A production, planner-routed multi-model agent (Claude + Amazon Nova) embedded in the MES, answering natural-language questions over live production data and guiding staff in-product — collapsing a week of cross-team investigation per customer concern.
The React 19 / MUI / Vite design system used across the application portfolio — token-driven light/dark theming, a layout shell, and a reusable component kit (command palette, data grid, dialogs, status chips, empty states) — so a new app starts from a working, consistent shell instead of a blank page. Now behind 15+ real-time operational dashboards.
The serverless analytics behind a 3,500+ device fleet dashboard — utilization and idle detection over a state engine, a unified per-device 360 view, and Zendesk / Pega service-desk integrations — unifying five sites and nine production lines. Later moved its endpoints off full-table scans onto precomputed rollups with async backfill.
The device-side edge of ingestion: getting heterogeneous machines to speak in the first place — an MQTT topic taxonomy, connectivity and heartbeat monitoring, Kafka consumers, and normalization of vendor-specific machine output into one common event shape, with faults routed as they happen. It gave production facilities real-time fault visibility for the first time.
A single reusable Lambda authorizer validating Cognito and Azure AD (Entra) JWTs plus scoped API keys across HTTP, REST, and WebSocket APIs, with a dynamic RBAC engine rolled out log-then-enforce to zero mismatches — standardizing secure access across internal operational applications.
Turned alerts into one-click action: Adaptive Cards that fire into Microsoft Teams on threshold breaches and failure-rate spikes, with operator buttons calling back into Lambda to remediate the device directly from chat — no console, no runbook hunt.
In-browser rendering of dental 3D meshes (STL / PLY / OBJ) with reference-vs-test geometric deviation heatmaps and PDF / zip export for quality control.
I started with computer science fundamentals, then moved into production engineering across cloud platforms, telemetry, dashboards, integrations, AI workflows, and deployment automation.
Glidewell Dental · Irvine, CA
Software Engineer III (2025 – Present) · Software Engineer II (2022 – 2025) · Software Engineer I (2021 – 2022)
Joined as the first software engineer on a hardware engineering team bringing connected devices online across five sites, and built the cloud telemetry backend and the application platform above it. Own architecture and delivery end to end — design the systems, deploy the infrastructure, then guide two engineers through implementation and review changes before production.
Bachelor of Science in Computer Science
Python, JavaScript / Node.js, SQL
React 19, MUI, Vite, Next.js, Tailwind, Plotly, Highcharts, WebSocket / MQTT clients
Lambda, IoT Core, API Gateway (REST / HTTP / WS), DynamoDB, RDS, S3, SQS / SNS / SES, EventBridge, Athena, MSK, Cognito, EC2
Amazon Bedrock (Claude, Nova), agent orchestration & tool use, NL-to-SQL, prompt engineering, human-in-the-loop design
DynamoDB data modeling, PostgreSQL, Athena, InfluxDB, Grafana, CloudWatch, MQTT pub/sub, Kafka consumers
Terraform, GitLab CI/CD, Jenkins / CloudBees, IAM, Entra ID (MSAL), OAuth 2.0 / JWT, RBAC, X.509 / mTLS
Contact
I am open to senior full-stack, cloud-platform, and event-driven systems roles — where AWS architecture, infrastructure leverage, real-time data, frontend standards, and production AI matter.
devnandol@gmail.com Irvine, CA (551) 234-1189