Blake Martin
Machine Learning Engineer
Building production ML systems that scale. Specializing in intelligent automation, multi-agent architectures, and MLOps pipelines that drive real business impact.
Technical Skills
Machine Learning
MLOps & Production
Data & Vector
AI Frameworks
Impact
Experience
Data Engineer
CM Incorporated
Engineered a manufacturing planning system with FastAPI and PostgreSQL, integrating real-time data pipelines and custom job grouping heuristics for dynamic scheduling (80% time savings). Built N8N-based automation agents with human-in-the-loop workflows across Sheets, payroll, and email. Developed robust data transformation pipelines for production metrics (e.g., fiber variance, roll weight, throughput), enabling downstream ML integrations and forecasting prototypes.
Key Achievements:
- • Developed custom optimization algorithms reducing operational costs by $100K+ annually
- • Implemented real-time data pipeline processing 10K+ manufacturing events daily
- • Built automated quality control system using statistical process control methods
- • Designed fault-tolerant job scheduling system with 99.9% uptime SLA
Machine Learning Engineer
AIC Incorporated
Built production ML systems including XGBoost sales forecasting model enhanced with GPT-3.5-turbo synthetic feature generation, achieving R² = 0.88 and 22% revenue boost. Developed automated hiring pipeline combining GPT-4 analysis with BiLSTM sentiment classification, achieving 91.4% accuracy in candidate evaluation and reducing screening time by 60%.
Technical Contributions:
- • Built XGBoost sales forecasting with GPT-3.5-turbo feature engineering (R² = 0.88, 22% revenue increase)
- • Developed GPT-4 + BiLSTM automated hiring system with 91.4% candidate evaluation accuracy
- • Implemented MLflow tracking and AWS Sagemaker deployment for production inference
- • Designed A/B testing framework for model performance evaluation and comparison
- • Established automated retraining pipelines with performance monitoring and drift detection
Machine Learning Intern
AIC Incorporated
Developed multi-agent LangChain system for automated scholarship essay evaluation, combining semantic search with Weaviate vector database and GPT-2 fine-tuning for personalized feedback generation. Led cross-functional team of 5 members in building agentic workflow that analyzed 2,000+ scholarship essays, helping 50+ students secure $120,000+ in scholarships through AI-powered essay improvement recommendations and automated scoring.
Project Outcomes:
- • Built multi-agent architecture with semantic similarity search for essay evaluation
- • Fine-tuned GPT-2 model (124M parameters) for domain-specific feedback generation
- • Implemented vector database with 2,000+ embedded scholarship essays for contextual analysis
- • Deployed inference pipeline on HuggingFace with Gradio interface for user interaction
- • Achieved 87% accuracy in essay quality prediction and recommendation relevance
Key Projects
Intelligent manufacturing planning system with custom ML-driven partitioning algorithms and real-time optimization. Built automated workflow orchestration processing 1K+ orders monthly, achieving 40% waste reduction and $100K+ annual cost savings through advanced job grouping algorithms.
Technical Highlights:
- •Custom partitioning algorithms for job optimization
- •Real-time data pipeline with queue management
- •Automated manufacturing workflow orchestration
Production email automation platform using LangGraph multi-agent workflows for intelligent email classification, prioritization, and response generation. Implemented vector-based memory retrieval with Pinecone and cost-optimized inference by hosting DeepSeek on RunPod, achieving 94.2% classification accuracy and processing 150 emails/minute with <2.5s response times.
Technical Highlights:
- •Multi-agent workflow orchestration with LangGraph
- •Vector memory system for contextual responses
- •Cost-optimized inference with DeepSeek on RunPod
- •Real-time email processing at scale
Personal project building sophisticated email intelligence system combining RoBERTa sentiment analysis, UMAP+HDBSCAN clustering, and predictive analytics. Features multimodal processing with BLIP/CLIP vision models, semantic search with FAISS/Weaviate, and agent-based architecture processing 5M+ emails.
Technical Highlights:
- •Multimodal content processing with vision transformers
- •Advanced clustering with UMAP dimensionality reduction
- •Real-time sentiment analysis and predictive timing
What Colleagues Say
"Blake consistently broke down complex problems into manageable parts and delivered thoughtful, efficient solutions. His accountability, attention to detail, and drive to improve made a real impact on both our operations and team culture."
Kevin Yen
CEO & Co-Founder, FlowData.ai
"Blake led cross-functional teams, built engaging educational products, and consistently pushed to improve our programs and processes. He's sharp, driven, and brings a contagious energy to the team."
Alex Duffy
Head of AI, Every Inc.
Machine Learning Newsletter
Weekly insights on emerging AI technologies. Join 2,600+ ML engineers following my technical deep-dives.
Beyond the Code: MCP — The End of AI Integration Hell
Technical deep-dive into Model Context Protocol and how it's solving the fundamental challenge of connecting LLMs to production systems. Covers architecture patterns, implementation details, and real-world tradeoffs.
Featured: MCP Architecture, JSON-RPC Implementation, Production PatternsGoogle's Prompt Engineering Bible & Hugging Face's Agents
Analysis of Google's prompt engineering best practices and Hugging Face's new agent frameworks. Practical insights for building reliable LLM applications.
Featured: Prompt Engineering, Agent Frameworks, Production LLMsOpenAI's Million-Token GPT-4.1 & Google's A2A Protocol
Technical breakdown of OpenAI's extended context windows and Google's Agent-to-Agent communication protocol. Implications for large-scale AI system architecture.
Featured: Extended Context, A2A Protocol, System ArchitectureLet's Build Something Great
I'm always excited to discuss ML engineering challenges, system architecture decisions, or opportunities to build scalable AI systems that drive real business value.
Get In Touch