Description

8 years of experience delivering enterprise-grade solutions across Data Engineering, Machine Learning, Natural Language Processing (NLP), Computer Vision, LLM’s and Generative AI solutions. I bring a unique combination of analytical rigor, system design expertise, and domain versatility. I have successfully delivered impactful solutions across industries including telecommunications, education, energy, manufacturing robotics, and cybersecurity.

I have designed and deployed scalable ETL/ELT pipelines, real-time data streaming systems, and cloud-based data lakehouse architectures. My work often bridges the gap between engineering and data science building robust infrastructure for training and serving ML models, and enabling seamless integration of NLP and computer vision pipelines into production environments. I bring deep expertise in Python, SQL, Apache Spark, Airflow, BigQuery, AWS/GCP, TensorFlow, PyTorch, and containerized deployments.

Whether building a semantic similarity engine using BERT, developing a Mask R-CNN model for agricultural yield estimation from drone imagery, or constructing multi-modal LLM pipelines for document parsing and translation, I focus on end-to-end value—from raw data ingestion to business-facing insights.

I look forward to bringing this impact-driven mindset and technical depth to your team.

Languages

English
Native or bilingual

Workplace preferences

Remote only

Primarily works remotely

Bespoke labs AI
Lead Machine Learning Engineer / AI
TECH
April 2026 - Today (2 months)
Mountain View, United States
Develop systematic strategies and recipes for creating high-quality RL environments that effectively train and evaluate agents.

Study how LLMs and agents fail across different task types, identifying patterns that inform better environment design.

Create benchmark environments that test specific agent capabilities, packaging them for external release on our evaluation platform.

Verify environment quality through hands-on testing—training small-scale agents, checking for reward hacking, and analyzing training dynamics.

Work with our environment creation pipeline to scale production of validated environments.
Machine learning Python Agents
RIOS Intelligent Machines
Senior Machine Learning Engineer
November 2024 - July 2025 (8 months)
Palo Alto, CA, USA
Redesigned Video Processing Pipelines: Built a scalable, reproducible ML workflow for anomaly detection in robotics using Metaflow and YOLO-based models.
Redesigned Video Processing Pipelines: Implemented parallelized frame filtering with support for Kubernetes-based scaling and integrated real-time logging for monitoring and diagnostics.

Engineered Dynamic Dataset Management System: Developed advanced data loaders and sequencing tools to preserve temporal and spatial relationships in multi-camera robotic video feeds.
Engineered Dynamic Dataset Management System: Enabled frame-level organization and retrieval by sequence, camera, and metadata attributes, boosting model training effectiveness on time-sensitive data.
Integrated FiftyOne and Encord in ML Loop: Built a full-cycle data-to-annotation-to-model pipeline, automating dataset filtering, annotation transfer, and model retraining.
Integrated FiftyOne and Encord in ML Loop: Implemented video stitching, annotation import/export, and dynamic dataset versioning through “Thoth v0” architecture.
Designed Modular Computer Vision Operators: Developed standardized ML pipeline components, including Apply Model, Segmentation Training, and Annotation Sync operators.
Designed Modular Computer Vision Operators: Made operators accessible via both GUI and API to support cross-functional robotics teams.
Computer Vision Deep Learning Machine learning Data Engineer Python
Unstructured IO
Senior Machine Learning Engineer
April 2023 - April 2025 (2 years)
San Francisco, CA, USA
Spearheaded comprehensive benchmark study of 10+ Vision-Language Models (Claude 3.5, OpenAI GPT-4o/3.5, Gemini Pro), improving table structure recognition by 15% and boosting image-based text extraction by 20% through model selection optimization.
Designed and deployed scalable multi-agent orchestration system using LangChain, AutoGen, and Pydantic AI, enabling collaborative LLM reasoning and reducing enterprise document processing time by 45%.
Built end-to-end RAG pipeline integrating layout detection, PaddleOCR/Tesseract OCR, and PDF parsing, achieving 30% faster throughput across enterprise workflows.
Fine-tuned transformer-based OCR models using LoRA on 11,000+ domain-specific technical PDFs, improving text accuracy by 12% and reducing missing text cases by 15% across mission-critical enterprise systems.
Machine learning Computer Vision Data Engineer Python Software Engineer