Software Engineer Machine Learning

Richard Brawley

*******.************@*****.*** +1-404-***-****

LinkedIn League City, TX 77573

Summary

Accomplished software engineer with 11+ years of experience leading full-stack AI/ML development in startup and enterprise environments. Expert in building and deploying large-scale NLP, LLM, and computer vision systems on cloud platforms (AWS, GCP, Azure) with modern DevOps practices (Docker, Kubernetes, CI/CD). Proven track record of driving product vision and scalability—spearheading AI initiatives at startups and delivering secure, HIPAA-compliant solutions in healthcare. Specialized in LLM orchestration and multi-agent architectures using LangChain and LangGraph. Skilled in translating complex requirements into technical architectures, mentoring engineering teams, and delivering high-impact AI products end-to-end.

Skills

•Programming & Frameworks: Python (FastAPI, Django, Flask), JavaScript/TypeScript (Node.js, React), Java, C++, Matlab, Go, gRPC, MCP

•AI & Machine Learning: TensorFlow, PyTorch, Keras, scikit-learn, Hugging Face Transformers, spaCy, OpenCV, YOLOv5, LangChain, LangGraph, Multi-Agent Systems

•LLM & GenAI: GPT-4, Retrieval-Augmented Generation (RAG), Prompt Engineering, PGVector, Haystack, Conversational AI, Autonomous Agent Design

•Cloud & DevOps: AWS (EC2, S3, Lambda, SageMaker), GCP, Azure, Docker, Kubernetes, Terraform, CI/CD (Jenkins, GitHub Actions, GitLab CI, ArgoCD)

•Data & Databases: PostgreSQL, MySQL, MongoDB, Redis, Apache Spark, Hadoop, Pandas, NumPy

Experience

Senior Software Engineer – Future Mind Labs, LLC (Jan 2023 – Present)

•Cognify (Custom LLM Query System):

oArchitected and implemented Cognify, an LLM-powered query interface for building data using FastAPI and the GraphQL API. Integrated OpenAI GPT-4 to translate natural language prompts into precise GraphQL queries, reducing query development time and eliminating syntax errors.

oDeveloped the retrieval pipeline using PGVector and Haystack: ingested building metadata and query templates into PostgreSQL with vector embeddings for retrieval-augmented generation (RAG). Achieved sub-second response latency and improved query relevance for complex data retrieval tasks.

oIntegrated a conversational UI using LibreChat directly into GraphQL Explorer console. Enabled in-situ query formulation and immediate result visualization, boosting customer usage and accelerating data access workflows.

oContainerized backend microservices and used gRPC for inter-service communication to ensure scalability and reliability. Deployed on Kubernetes with auto-scaling, the platform sustained thousands of concurrent sessions at 99.9% uptime and handled surge in query traffic without performance degradation.

•FuAssistX (Smart Multi-Tool AI Assistant):

oDesigned and built FuAssistX, a multi-tool AI chatbot powered by OpenAI GPT-4 and the Model Context Protocol (MCP). Created FastAPI microservices that expose domain-specific functions (e.g., data lookup, analytics) as MCP tools, enabling the LLM to invoke external services (via gRPC) on demand.

oLeveraged FastMCP to orchestrate secure connections between the chatbot and data sources (GraphQL, REST APIs). Configured the MCP server to allow the LLM to invoke tools dynamically, enhancing the chatbot’s ability to handle complex queries and reducing response errors.

oDesigned a multi-agent architecture using LangGraph, enabling GPT-4 to act as coordinator across autonomous tools for dynamic task execution (data retrieval, analytics, summarization). This agentbased approach increased the modularity and reliability of the system.

oImplemented a hybrid knowledge retrieval pipeline using Haystack and PGVector: vectorized company documents and FAQs in PostgreSQL to power retrieval-augmented answers. This enabled FuAssistX to provide contextually accurate information, improving answer accuracy and covering a broader range of domain knowledge.

oEngineered the full-stack infrastructure (FastAPI, gRPC, Docker, Kubernetes) and integrated LibreChat for the user interface. Delivered a seamless chat experience with sub-300ms median response times.

Senior Machine Learning Engineer, Integral Ad Science (Sep 2021 – Dec 2022)

•Led design and implementation of scalable NLP data pipelines (tokenization, lemmatization) on Databricks for the AdContext NLP Optimizer project, enabling downstream webpage classification, sentiment analysis, and entity recognition; leveraged PostgreSQL for structured text storage.

•Developed and fine-tuned transformer-based NLP models (BERT embeddings) on AWS (EC2/SageMaker) to enhance content relevance prediction, boosting classification accuracy up to 90% and driving around 15% uplift in targeted ad click-through rate (CTR).

•Utilized MLflow for experiment tracking and Databricks notebooks for collaborative development, orchestrating reproducible end-to-end ML workflows from data ingestion through model versioning.

•Deployed production-grade FastAPI microservices (Docker containers) on AWS (ECS/EKS), integrating Redis caching and PostgreSQL to achieve sub-100ms inference latency and support high-throughput real-time ad serving.

•Collaborated with data science, engineering, and ad-ops teams to integrate NLP-derived context signals into adserving pipelines, improving ad relevance and placement accuracy, while enhancing overall campaign ROI.

•Monitored key performance metrics (accuracy, latency) in production via AWS CloudWatch and custom dashboards; proactively retrained and optimized models to maintain over 90% accuracy and sub-100ms inference latency under dynamic load.

Full Stack AI Engineer, MediHealth AI, Inc (Apr 2020 – Aug 2021)

•Ensured HIPAA compliance in all development processes by implementing encrypted data pipelines and secure storage for protected health information (PHI), adhering to strict healthcare data privacy regulations.

•Developed computer vision models using TensorFlow CNNs to analyze medical imaging (X-rays, MRI), achieving ~90% accuracy in detecting diagnostic anomalies and accelerating clinical decision support.

•Built NLP workflows with BERT-based Transformers to extract key insights from electronic health records and clinical notes, improving efficiency of patient risk stratification and reporting.

•Engineered full-stack applications (React.js front-end, Node.js/Flask back-end) for patient monitoring dashboards, integrating AI predictions and real-time data visualizations to support healthcare providers.

•Deployed AI services on AWS and Azure with secure VPC configurations and automated scaling. Established containerized environments (Docker, Kubernetes) and GitLab CI pipelines for continuous delivery of ML models, ensuring reliable and scalable production deployments.

•Collaborated closely with clinicians and domain experts to validate model outputs and incorporate medical protocols, ensuring AI solutions met healthcare standards and delivered actionable results.

AI/ML Engineer, GlobalTech Corp (Aug 2018 – Mar 2020)

•Developed predictive ML models using Python (scikit-learn, TensorFlow/Keras) for image and signal data analysis, achieving over 90% accuracy on classification tasks and improving key performance metrics.

•Created RESTful APIs and web interfaces (Flask/Django back-end, React front-end) to serve ML models and present interactive data visualizations, enabling stakeholders to access AI-driven insights easily.

•Automated data preprocessing and feature engineering pipelines to handle large-scale datasets, incorporating data cleaning, augmentation, and efficient batch processing to enhance model training.

•Maintained and improved DevOps pipelines using Docker and Jenkins, and implemented monitoring with Prometheus/Grafana to ensure high uptime and performance of production AI services.

•Researched and prototyped advanced AI techniques (GANs, transfer learning) to enhance model robustness and address evolving project requirements, contributing to the company’s innovation roadmap.

Machine Learning Engineer, Becton, Dickinson and Company (Mar 2016 – Jul 2018)

•Designed and implemented deep learning models for medical device diagnostics. Developed CNN-based image classification models (TensorFlow/Keras) to analyze clinical imagery, achieving high accuracy in detecting anomalies. Implemented real-time object detection using YOLO to identify issues in lab instrumentation, reducing manual review time by over 20%.

•Built natural language processing pipelines to analyze clinical reports. Used NLTK and spaCy to extract key information from patient and lab records, automating feature extraction and streamlining data analysis.

•Architected cloud-based data pipelines on GCP: managed research and imaging data with BigQuery, automated preprocessing tasks using Cloud Functions, stored large datasets on Google Cloud Storage, and deployed ML workflows on Vertex AI and Kubernetes Engine. This infrastructure enabled scalable model training and efficient production deployment of AI solutions.

Machine Learning Intern, Houston Mechatronics (May 2015 – Feb 2016)

•Developed image analysis and machine learning prototypes for diagnostic devices. Used ImageJ and MATLAB for medical image preprocessing and segmentation, applied principal component analysis (PCA) for feature extraction, and implemented prototype CNN models (TensorFlow/Keras) to classify sample data.

•Automated data processing tasks in Python to accelerate experimental analysis. Created scripts to clean, transform, and visualize new datasets, which reduced manual preparation time and enabled faster iteration on ML algorithms. Documented findings and collaborated with senior engineers to refine models and improve overall prediction accuracy.

Education

•Master of Science in Computer Science, Rice University Houston, 2015

•Bachelor of Science in Computer Science, The University of Texas at Austin, 2013

Contact this candidate