R

Robert robert Chilton

About

Detail

McKinney, Texas, United States

Timeline


work
Job
school
Education

Résumé


Jobs verified_user 0% verified
  • Lightedge
    Senior Software Engineer - AI/ML
    Lightedge
    Mar 2020 - Oct 2025 (5 years 8 months)
    • Led design and delivery of a multi-tenant NLP/GenAI platform on AWS, centered on PyTorch models fine-tuned with PEFT/LORA for classification, Q&A, and summarization; trained on SageMaker with spot instances and data in S3. • Built RAG pipelines using OpenSearch k-NN and FAISS for semantic retrieval; implemented chunking/embedding strategies, hybrid BM25+ANN search, and per-tenant prompt templates with guardrails. • Stood up high-throughput inference via TorchServe on EKS for encoder/seq2seq workloads and vLLM for LLM decoding (2023+), with token-level batching and autoscaling tied to queue depth. • Integrated managed foundation models where appropriate via Amazon Bedrock (GA 2023) alongside first-party PyTorch models; routed traffic by ta
  • Appfolio
    Senior ML/AI Engineer
    Appfolio
    Feb 2018 - Jan 2020 (2 years)
    • Built text classification and NER for property-management communications (work orders, tenant messages) using BERT-family encoders (2019+) in PyTorch, with Hugging Face Transformers; earlier models leveraged fastText embeddings and scikit-learn baselines. • Productionized training/inference on SageMaker (training jobs, Batch Transform, hosted endpoints) with data in S3; orchestrated ETL and retraining with Airflow and event-driven Lambdas. • Introduced MLflow for experiment tracking and model comparison; defined dataset versioning and evaluation suites to guard against label drift. • Delivered semantic search over maintenance logs using FAISS for ANN retrieval and BM25 for lexical fallback. • Exposed models through small Flask/FastAPI ser
  • Blackbaud
    Mid Data & ML Engineer
    Blackbaud
    Jan 2017 - Jan 2018 (1 year 1 month)
    • Prototyped donor-engagement scoring with PyTorch RNN/CNN text models and spaCy/scikit-learn baselines; productionized the best model for batch scoring of campaign notes and messages. • Built data preparation pipelines for text normalization, PII scrubbing, and labeling; packaged batch inference via AWS Batch and EC2, storing outputs in S3 for downstream reporting. • Implemented simple NER with spaCy to enrich constituent profiles; validated gains against held-out campaigns. • Served inference behind a lightweight Flask API for internal tools (kept backend surface area intentionally small).
  • Granicus
    Junior Software Engineer / Internal Tools
    Granicus
    Jun 2015 - Dec 2016 (1 year 7 months)
    • Built Flask admin tools with SQLite to orchestrate simulation scenarios and export telemetry; added RBAC/JWT and audit hooks. • Introduced CI for mixed Windows/Linux stacks with GitHub + Jenkins, PowerShell/Bash, and authored pytest suites.
Education verified_user 0% verified
  • T
    Bachelor of Science in Computer Science
    Texas A&M University-San Antonio
    Jan 2011 - Jan 2015 (4 years 1 month)
    Focus Focused on Distributed Systems, Scalable MLArchitectures, and Cloud Engineering. Relevant Coursework: Data Structures and Algorithms, Object-Oriented Programming, Operating Systems, Computer Networks, Database Systems, Software Engineering, Artificial Intelligence, Computer Graphics.