Work Experience
Amazon WW Grocery
- •HAWK: Predictive Operations Intelligence — Multi-site operational intelligence platform with subprocess-level bottleneck detection, comparative performance benchmarking, and predictive ETAs.
- •Architected hybrid system (SNS/Lambda, time-series + NoSQL, SageMaker + LLM) achieving sub-200ms latency, reducing MTTR by ~35% and process variance by 40%.
- •Built time-series analytics engine surfacing cyclical patterns; increased manager adoption from 30% to 85% and reduced unplanned overtime by 25%.
- •Grocery Identification Microservice — Owned end-to-end design of near real-time event processing service consuming 10K-100K inventory updates/day.
- •Built scalable queue-based infrastructure with serverless compute and retry logic, enabling 6+ services to consume events with zero code changes.
- •Led 3 engineers through system design reviews, implementation, code reviews, automated testing, deployment, and monitoring.
- •Multi-Region Data Migration — Re-architected inventory datastore from relational to NoSQL (DynamoDB), achieving 10x throughput scalability while reducing infrastructure costs by 10%.
- •Designed and executed zero-downtime migration of 10M+ records with < 5 minutes downtime and comprehensive validation across multiple regions.
Amazon Physical Stores
- •Cloud & Infrastructure Migration — Migrated high-traffic backend microservice to serverless Lambda + API Gateway architecture, reducing infrastructure costs by 50%.
- •Developed migration scripts for 100+ fulfillment centers worldwide, orchestrating SKU data transfer and enabling full deprecation of legacy warehouse management system.
- •Tool Modernization & Performance — Led migration of 3+ internal tools from AngularJS to React, designing incremental rollout with feature flags for safe, reversible production deployments.
- •Optimized web application performance by 49% (5s to 2.4s) through bundle analysis and dependency optimization, saving ~$68 per user in operational costs.
Work Summary
Deep dives into systems I designed and shipped — architecture, trade-offs, and impact.
Barcode Linked to Multiple ASINs
Automatically transfers on-hand inventory from a discontinued product to its replacement when the catalog merges them.
- projected annual savings
- $1.55M
- stores
- 600+
- reconciliation, from thousands of hours
- < 24h
- SNS
- SQS
- DynamoDB
- AWS Batch
- S3
HAWK
Multi-site operational intelligence platform with subprocess-level bottleneck detection, cross-site benchmarking, and predictive ETAs.
- lower MTTR
- 35%
- less process variance
- 40%
- manager adoption
- 30→85%
- AWS Lambda
- SNS
- SageMaker
- DynamoDB
- Java
Grocery Identification
Near real-time event processing layer that separates grocery items from Amazon Core Inventory for downstream consumers.
- events / day
- 10K–100K
- services onboarded
- 6+
- consumer code changes
- 0
- SNS
- SQS
- AWS Lambda
- DynamoDB
- Kotlin
Multi-Region Data Migration
Re-architected an inventory datastore from relational to DynamoDB and migrated it across multiple regions.
- throughput scalability
- 10x
- records migrated
- 10M+
- downtime
- < 5 min
- DynamoDB
- AWS
- Multi-region
- Data migration
AI & Automation Interest
Growing interest in exploring how AI and automation can enhance operational systems, developer productivity, and system intelligence.
LLM Integration
Integrating Claude, GPT models, and other LLMs into systems for intelligent automation and decision-making
Workflow Automation
Building automated systems using modern AI tools to reduce manual toil and improve operational efficiency
Intelligent Agents
Designing agentic systems that can make decisions, take actions, and adapt based on operational context
AI-Driven Operations
Applying AI to operational intelligence, predictive analytics, anomaly detection, and system optimization
Education
M.S., Computer and Information Sciences
Gainesville, FL