Senior AI/ML Engineer with 7 Years Experience

Dennis William Mitchell

Senior AI/ML Engineer

****.**@*******.*** Staten Island, NY

SUMMARY

Senior AI/ML Engineer with about 7 years of experience across retail logistics, fintech, and healthcare tech. I'm someone who actually enjoys the collaborative side of engineering - whether that's sitting down with a product manager to untangle vague requirements, or helping a junior engineer debug a Kafka consumer that keeps falling behind. I'm very result-focused: I'd rather ship a simple, working feature that users actually need than spend months polishing something nobody asked for. Over the years, I've learned that the best technical decision is usually the one that unblocks the team, not the one that's most elegant. SKILLS

LLMs & Generative AI: GPT-4/4o, Claude 3.5 Sonnet, Gemini, AWS Bedrock, Prompt Engineering, RAG, Fine- tuning, Guardrails

Agentic AI: LangChain, LangGraph, AWS Bedrock Agents, Tool Calling, Multi-Agent Workflows, Memory Management

Vector Databases: Pinecone, Qdrant, Milvus, FAISS, pgvector, Azure AI Search ML & Deep Learning: PyTorch, TensorFlow, Hugging Face Transformers, Scikit-learn, XGBoost, spaCy, OpenCV MLOps & Experiment Tracking: MLflow, Weights & Biases, Kubeflow, Feast, Model Registry, BentoML Backend & APIs: Python (FastAPI, Flask), REST, GraphQL, WebSockets, gRPC Cloud Platforms: AWS (SageMaker, ECS, EKS, Lambda, Bedrock), Azure (AKS, OpenAI, AI Search), GCP (Vertex AI) DevOps & Infrastructure: Docker, Kubernetes (EKS/AKS), Terraform, Helm, GitHub Actions, ArgoCD Data & Streaming: PostgreSQL, Redis, Kafka, RabbitMQ, Airflow, Spark Version Control: Git (trunk-based, bisect, reflog), GitHub Actions, conventional commits Observability: Prometheus, Grafana, OpenTelemetry, LangSmith, Arize Phoenix EXPERIENCE

Senior AI/ML Engineer, Helpware Tech - San Francisco, CA 03/2022 – 03/2026

•Collaborated with a team of four engineers to build a generative AI assistant for internal customer support using LangChain, Pinecone, and GPT-4. Went from whiteboard to MVP in about three months.

•Helped design a RAG pipeline for knowledge base retrieval – document chunking, embedding generation via AWS Bedrock Titan, and vector storage in Qdrant.

•Built and maintained model inference services using FastAPI, deployed as Docker containers on AWS ECS with autoscaling based on request volume.

•Set up Git LFS for model weight files and Terraform state. Contributed to Helm charts for deploying inference pods to production EKS clusters.

•Participated in establishing a trunk-based Git workflow with short-lived branches and required PR approvals.

•Mentored two junior engineers on integrating LLM APIs into backend services and debugging token limit issues with streaming responses.

•Responded to a production incident when a model started returning poor results. Used git bisect to find the offending commit, rolled back, and restored stability within an hour.

•Established a trunk-based Git workflow with short-lived feature branches (max two days), automated semantic releases using semantic-release, and required PR approvals from at least two people. New hires could make their first commit without breaking anything.

AI/ML Engineer, Computools - New York, NY 09/2020 – 02/2022

•Transitioned into AI/ML gradually. Started by contributing to a recommendation feature for an e-commerce dashboard, then helped maintain a document intelligence product.

•Contributed to a document parsing platform: FastAPI backend calling a fine-tuned LayoutLMv3 model, PostgreSQL schemas for extraction results, and an Angular grid for user review.

•Helped set up GitHub Actions CI/CD pipelines – linting, tests, Docker builds, deployment to AWS ECS on merge to main.

•Wrote reusable Terraform modules for AWS resources (S3 event triggers, RDS clusters) adopted by another team.

•Integrated real-time sentiment analysis into an Angular dashboard using WebSockets streaming predictions from a PyTorch BERT model on Kubernetes.

•Used MLflow to track training runs – hyperparameters, metrics, artifacts – tied to Git commit hashes for reproducibility.

Software Developer, End Point Dev - New York, NY 07/2018 – 08/2020

•Started as a generalist on a seven-person team building internal tools for data operations. Worked on backend APIs, small frontend components, and DevOps tasks.

•Helped build an Angular dashboard displaying outputs from a prototype computer vision model used daily by the data science team.

•Dockerized a labeling tool on Flask and deployed to AWS EC2, reducing new user setup time from a day to about fifteen minutes.

•Worked with a senior engineer to refactor legacy code into modular, testable functions.

•Helped two interns get their first PRs merged – resolving merge conflicts, using git stash, writing clear commit messages.

CORE PROJECTS

•LLM Meeting Summarizer with RAG – Built a tool using LangChain, GPT-4, and vector search for meeting transcript context. Deployed on AWS ECS with Docker and GitHub Actions. Worked with PM to focus on fast, accurate summaries over fancy features.

•Visual Search for Retail – Built image upload, ResNet50 embeddings, Milvus vector search, and FastAPI backend. Deployed on Kubernetes with Terraform.

•Real-time Anomaly Detection – Built XGBoost-based detection for manufacturing sensor data with Kafka streaming, Redis alerts, and PostgreSQL logging.

•LLM Playground for Engineers – Built a sandbox for testing Hugging Face models with adjustable parameters. Backend used FastAPI and Ray for distributed inference. Used by 30+ engineers.

•Customer Feedback Classifier – Built DeBERTa-based sentiment and urgency classification. Deployed as Docker container on AWS Lambda behind API Gateway. EDUCATION

CUNY, College of Staten Island, Bachelor's Degree in Computer Science 08/2014 – 05/2018 Staten Island, NY

Contact this candidate