Senior AI/ML Software Engineer (RAG & Graphs)

Matthew Fernandez

Senior Software Engineer

Leesburg, VA 20176 442-***-**** **********@*****.*** Linkedin SUMMARY

Senior Software Engineer with 10+ years of experience building and deploying scalable machine learning and LLM-powered systems, specializing in RAG architectures, knowledge graphs, semantic search, and cloud-native solutions across AWS and Azure.

PROFESSIONAL EXPERIENCE

Palantir Technologies Leesburg, VA

Senior AI Engineer 04/2013 – 06/2026

● Developed and architected a production-ready multi-tenant Knowledge Graph REST API using Python, FastAPI, and Neo4J to manage complex asset hierarchies and performance KPIs for enterprise customers.

● Designed and implemented a scalable graph database architecture with full-text search capabilities, graph traversal algorithms, and intelligent relationship suggestions, enabling users to explore interconnected asset and performance data across multiple business contexts.

● Built a comprehensive API with role-based access control (RBAC), tenant isolation, background job processing for data import, and OpenAPI documentation, serving as the core data layer for asset management and performance analytics in a microservices environment.

● Implemented robust graph operations including multi-graph traversal, BM25 full-text search with relevancy scoring, similarity edge detection, and KPI hierarchy modeling, with comprehensive test coverage using BDD (Behave), integration tests (Tavern), and unit tests.

● Built scalable AI-powered video analytics pipelines on Databricks, reducing 26-minute videos to 2-minute actionable clips while enabling concurrent processing of 20+ videos using serverless GPU infrastructure.

● Delivered AI-driven financial data intelligence solutions that achieved 90% fraud detection accuracy while reducing operational costs by 34% through real-time analytics and machine learning.

● Designed and developed an enterprise AI copilot for industrial asset performance monitoring using LLM orchestration, integrating OpenAI GPT models via Azure OpenAI with intelligent query routing to specialized function studios for data analysis, KPI impact assessment, ARIMA forecasting, and vector database semantic search.

● Implemented end-to-end document ingestion, embedding, and semantic retrieval pipelines using Hugging Face Transformers and internal vector databases, indexing millions of documents with sub-second query latency.

● Developed enterprise AI chatbot systems that combine LLM reasoning, retrieval augmentation, and structured tool calling, allowing internal teams to automate workflows and access domain-specific knowledge safely.

● Designed and implemented a graph-based maritime trajectory analytics pipeline that ingests AIS data, preprocesses vessel movements, and stores relationships in Neo4j AuraDB for semantic and geospatial querying.

● Built custom algorithms to chunk vessel trajectories, create Next voyage links, and categorize navigation attributes

(SOG/COG) to improve temporal path reconstruction and analytics accuracy.

● Designed a petabyte-scale streaming data ingestion pipeline that sustained 12GB/s throughput, processed 1PB of data and 1.04 trillion records in 24 hours using serverless, auto-scaling infrastructure.

● Developed geospatial utilities to identify closest ports, calculate distances, and update port and vessel metadata programmatically.

● Integrated AI embeddings for trajectory segments to enable intelligent similarity search and anomaly detection across maritime behavior patterns.

● Automated creation of Neo4j indexes, relationship modeling, and risk-property tagging enabling performance- optimized queries for large-scale maritime datasets.

● Processed multiple large AIS datasets (millions of entries) using Pandas, memory-efficient chunking, and multiprocessing methods to ensure scalability.

● Built a LangGraph/LangChain RAG system integrating Milvus vector search with local Ollama LLM, enabling natural- language querying and retrieval of maritime AIS trajectory data using custom prompt orchestration and document chunking.

● Implemented MLflow experiment tracking and optional Evidently data quality reporting to monitor model performance, context quality, and system metrics during retrieval and generation workflows.

● Built a full-stack geospatial search application integrating similarity search with OpenStreetMap services to retrieve and analyze overhead imagery across the San Francisco region using cosine-similarity vector queries and spatial indexing.

● Implemented an MCP server and modern front-end client interface to surface real-time search results via chat and map UI, demonstrating strong ability to architect scalable solutions, integrate third-party APIs, and communicate design decisions in technical discussions.

● Developed a client-side Kubernetes load balancing solution that reduced infrastructure costs by 20% through intelligent traffic distribution, improving latency stability and resource utilization.

● Built time-series forecasting models for sales, supply chain, and integrated business planning (IBP), benchmarking GRU, Transformer, and long short-term memory (LSTM) models within Apache Spark (Databricks).

● Implemented an AI assistant using reinforcement learning from human feedback (RLHF) and benchmarked it against PEFT using LoRA for optimized response quality.

● Developed and deployed an AWS SageMaker web page category classification model, improving content categorization accuracy by 20%, streamlining user navigation and content management.

● Implemented robust web scraping solutions from company websites and aggregators using Python and BeautifulSoup, achieving a 150% increase in data collection efficiency and providing critical insights for strategic decision-making.

● Built comprehensive monitoring and alerting infrastructure using Prometheus and Grafana with custom metrics collection, achieving 99.99% uptime SLA for production inference endpoints serving millions of users.

● Architected Kubernetes-based autoscaling and cost optimization system with custom scheduling policies, reducing infrastructure costs by 35% while maintaining strict performance SLAs.

● Implemented robust testing coverage including RSpec, Pytest, and Cypress for unit, integration, and E2E testing across multi-service domains, maintaining regression rate across releases with full CI/CD integration.

● Mentored engineers across squads, led cross-functional architecture reviews, and partnered closely with product, design, and SRE to align systems with user experience, performance, and reliability goals Blackbird Technologies Leesburg, VA

Software Engineer 07/2008 - 04/2013

● Built secure APIs and backend services compliant with HIPAA and other healthcare regulations, managing sensitive patient data and integrating with electronic health record (EHR) systems.

● Developed a synchronization trigger to channel data from PostgreSQL to Elasticsearch, increasing full-text search speed by nearly 90% and providing a list of suggestions for website search needs.

● Upgraded the Socket.IO module, optimizing it to handle multiple connections and user disconnects, which improved real-time communication reliability and boosted user satisfaction by 25%.

● Created scalable REST APIs and monitoring dashboards for real-time robot fleet status using React, Flask, and Grafana, improving incident triage time by 40%.

● Developed event-driven microservices for robot telemetry ingestion using Kafka and AWS Kinesis, enabling real-time monitoring of 100K+ robots across multiple FCs (fulfillment centers).

● Built internal simulation tools using Python and WebGL to visualize robot traffic, collision avoidance, and throughput bottlenecks, accelerating debugging and field validation. EDUCATION

The Johns Hopkins University 2004-2008

Bachelor of Science, Computer Science

SKILLS

Languages: Python, Go, Typescript, C#, SQL, Rust, Java, Scala Backend: FastAPI, Django, Flask, NestJs, Express.js, ASP.Net, Rails, Gin, Spring Boot Database: Postgresql, MSSQL, MongoDB, Elastic Search, Neo4J, Milvus, Redis, PostGIS, Pinecone AI/ML: LangChain, LangGraph, Pydantic, AutoGen, RAG, MCP, LLamaIndex, HuggingFace, Tensorflow, PyTorch, Transformer, Guardrails, rebuff, scikit-learn

Architectures: Distributed systems, Microservices, Modular Monoliths, Event-Driven Architecture, API-first Design(REST & GraphQL), Domain-Driven Design (DDD), CQRS, Scalable APIs, Multi-cloud orchestration, In-database ML pipelines ETL: AWS Glue, Amazon SageMaker, Azure Data Factory, Google VertexAI Big Data: Airflow, Snowflake, Apache Kafka, RabbitMQ, Flume, Hadoop, MapReduce, BigQuery, RedShift, Deequ Cloud & Infrastructure: Azure, Monitor, GCP, Firebase, Docker, Kubernetes, CI/CD, Jenkins DevOps: Git, Argo CD, Terraform, Swagger/OpenAPI, Ngnix, Grafana, Prometheus Testing: Jest, Pytest, Cypress, Playwright, Puppeteer Data Visualization: Leaflet.js, Mapbox, Openlayers, WebGL, Three.js, D3.js, Cytoscape, Pixi.js, GSAP, PowerBI Frontend: Next.js, React, Svelte, Angular, Vue, Streamlt Security & Compliance: Zero-Trust, OAuth2, RBAC, JWT, mTLS, GDPR, SOC2, PCI-DSS Collaboration & Leadership: Agile/Scrum, Mentorship, Code Reviews, Cross-team Collaboration, Technical Strategy, Architecture Reviews

Contact this candidate