I am an AI Technical Lead with a proven track record of delivering impactful business value through machine learning and computer vision solutions. Currently, I supervise PG&E's computer vision team as the Lead Machine Learning Engineer, where I lead 12 professionals—including data scientists, ML engineers, product managers, and data labelers—to build a system capable of inspecting PG&E’s entire grid in a single day. Previously, I led a team of 10 data engineers, driving key initiatives such as the analytics supporting PG&E’s $25 billion undergrounding project and fulfilling metric calculation requirements in compliance with California Senate Bill 884.
Key achievements in my career include:
I achieve results by fostering autonomy, mastery, and purpose within my team, creating psychological safety, and cultivating a feedback culture to drive continuous improvement. My approach builds trust and ensures high performance.
I excel at building strong stakeholder relationships by soliciting early and frequent feedback, tailoring communication to my audience, and promoting shared understanding. I believe data science is a team effort and look forward to collaborating on transformative initiatives.
Proficient in developing and deploying advanced AI solutions, including expertise in computer vision, tabular data models, and natural language processing. Skilled in Generative AI (GenAI) technologies and frameworks such as LangChain, Llama Index, Amazon Bedrock, OpenAI, Anthropic, Google Gemini, and Cohere. Extensive knowledge of methods and principles like Retrieval-Augmented Generation (RAG), prompt engineering, AI agents, and tool development. Demonstrated ability to design and implement AI systems that solve complex problems by leveraging classical machine learning techniques alongside state-of-the-art generative models. Strong focus on model monitoring, maintenance, and optimization to ensure sustained performance and reliability.
Experienced in designing scalable data pipelines and managing large-scale datasets, including structured, unstructured, and spatial data. Skilled in leveraging Palantir Foundry’s data management platform to enable effective data integration, transformation, and analytics. Proficient in database technologies, including Redshift, Teradata, Oracle, PostgreSQL, MySQL, and Microsoft SQL Server. Frequently develop finely tuned queries, ETL pipelines, and relational data models, implementing their physical representations to support efficient querying and analytical workloads.
Specialized in building production-grade and scalable machine learning systems, focusing on end-to-end pipelines from data preparation to model deployment. Proficient in designing systems that integrate seamlessly with cloud infrastructure and APIs. Experienced in implementing efficient training, validation, and inference pipelines that balance performance and scalability. Skilled in applying software engineering principles such as modular code design, version control, and CI/CD to ensure maintainability and reliability in machine learning workflows.
Extensive experience with AWS cloud services, including SageMaker, Lambda, S3, RDS, Rekognition, Textract, Ground Truth, IAM, IAM Identity Center, Step Functions, CodePipeline, CloudFormation, CodeArtifact, EFS, EC2, Elastic Beanstalk, SNS, CloudWatch, EventBridge, Secrets Manager, and more. Skilled in serverless architecture and containerized deployments leveraging tools like Docker and Kubernetes. Proficient in utilizing Infrastructure as Code (IaC) tools such as Terraform, Serverless Framework, and AWS CloudFormation to design and manage scalable, reliable, and repeatable cloud infrastructure.
Experienced in managing complex projects in agile environments, with expertise in Scrum and SAFe methodologies. Skilled in leading Scrum ceremonies, including daily stand-ups, sprint planning, refinement, retrospectives, and reviews. Comfortable managing multiple stakeholders, competing deliverables, and navigating cross-team dependencies. Adept at incorporating Kanban workflows where applicable and participating in Agile Release Train demo events to showcase progress and align priorities across teams.
Proven success in managing teams of up to 15 members while fostering collaboration and inclusivity. As the leader of PG&E’s Data Science Community of Practice (DSCoP), increased engagement by over 500%, achieving consistent membership of 70+ individuals. Developed coaching strategies tailored to individual needs, mentoring junior colleagues to advance their technical and professional growth. Spearheaded the creation of a 10-class Data Science Lifecycle Bootcamp, generating 20+ hours of instructional content and training over 200 employees. Previously taught Python programming to over 70 colleagues, promoting foundational skills development across diverse teams.
Experienced in strategic planning, resource allocation, and performance management. Skilled in providing constructive feedback to direct reports and aligning team efforts with organizational objectives. Adept at navigating complex internal business processes to ensure compliance and operational efficiency. Proven ability to manage cross-functional teams and maintain strong client and stakeholder relationships, delivering impactful results on time and within budget.
Exceptional communicator with a strong ability to present complex technical solutions to diverse audiences, including executive leadership, industry peers, and cross-functional teams. Delivered impactful conference presentations such as "Generative AI for Predictive Maintenance" at Utility Analytics in 2025, "Artificial Intelligence Enabled Aerial Inspections at PG&E" at the Utility Analytics Summit in 2020, and "Feature Engineering for Time-Series Electric Interval Usage Data" at the AEIC in 2018, which was the highest-rated presentation at the conference. Skilled in tailoring messaging to the audience, preparing detailed documentation, and facilitating technical training. Recognized for clear and engaging presentations that drive understanding and inspire action.
For a complete list of skills, please reference my LinkedIn.
May 2024 - Present
As Data Science Supervisor and Tech Lead, I manage a team of 12, including data scientists, ML engineers, product managers, and data labelers. I led the design and deployment of an automated computer vision system on AWS, which is capable of processing over 10 million images daily. Additionally, I directed the development of PySpark jobs that aggregate predictions into stakeholder reports. My responsibilities also include negotiating contracts, conducting team meetings, reviewing time entries, providing feedback, delivering training, and setting priorities and timelines for technical tasks.
July 2022 - April 2024
As the technical lead for both computer vision and data engineering teams, I managed 15 engineers and scientists. My data engineering team focused on driving analytics for PG&E's $25 billion undergrounding initiative, as well as developing risk-related metric calculations to ensure compliance with California Senate Bill 884. On the computer vision side, I led the development of synchronous and asynchronous model prediction REST APIs, utilizing AWS services such as API Gateway, Lambda, and SQS, along with technologies like Serverless, Jenkins, Postgres, and Kubernetes. Additionally, I designed and implemented a relational database for storing computer vision labels and developed PySpark data pipelines to process geospatial data using Apache Sedona on Palantir Foundry.
At Cornell, I focused on machine learning and software engineering.
March 2019 - October 2021
I contributed to an award-winning, agile team recognized by the Wall Street Journal for developing computer vision solutions to mitigate wildfire risk. I created PyTorch-based object detection and classification models to analyze drone, satellite, and helicopter imagery. Additionally, I developed a service to automate the training and deployment of object detection models using AWS GroundTruth annotations and built video segment classification models to identify key time spans in helicopter-captured imagery.
August 2020 - August 2021
In addition to my primary role as a Senior Data Scientist in machine learning, I led PG&E’s Data Science Community of Practice (DSCoP), a grassroots, company-wide initiative focused on advancing analytics knowledge sharing, community building, and standardizing data science practices. As DSCoP leader, I fostered an inclusive and psychologically safe environment, increasing active engagement by over 500% to a consistent membership of around 70 individuals. I led a team of data scientists in developing the Data Science Lifecycle Bootcamp, which included live instructional content, recorded videos, and coding exercises designed to teach best practices for each stage of the data science lifecycle in predictive analytics projects. More than 200 PG&E employees participated in the bootcamp.
January 2020 - August 2020
In addition to my role as Senior Data Scientist, I also served as Chief of Staff for my department. In this capacity, I managed administrative and leadership responsibilities, including organizing and facilitating all-hands meetings and leadership team discussions.
August 2017 - March 2019
I focused on customer-centric problems which required data mining over 1 trillion energy consumption records and building both semi-supervised and unsupervised learning models. During my tenure, I was able to present my research, "Feature Engineering for Time-Series Electric Interval Usage Data," at the Association of Edison Illuminating Companies conference in Lake Buena Vista, FL, and was honored as the highest-rated presenter. Additionally, I also taught an eight-week introductory Python class to over 70 of my colleagues at PG&E.
August 2015 - July 2017
I developed professional software engineering skills by collaborating with senior-level engineers to build RESTful web services. I was also able to work on a variety of interesting machine learning projects, learned the Spring and Django web frameworks, and challenged myself to enhance open-source code bases with functionality required for my team's projects.
June 2014 - August 2014
My time at Hyundai provided me with the opportunity to assess, develop, and implement solutions to a variety of complex challenges, such as optimizing supply-to-line vehicle routing and inventory levels. I also created a cost-savings strategy projected to save the company between $90,000 and $420,000 annually, while collaborating with teams across the plant.
While majoring in Industrial Engineering and Operations Research at the University of California, Berkeley, I participated in a range of enriching activities. I contributed to research alongside professors and graduate students, gained valuable industry experience through an internship with Hyundai, and led the Institute of Industrial Engineers. Additionally, I instructed Engineering 98, assisted with labs for CS 61B, and worked as a Residential Computing Consultant in the dorms.
At Cornell, I focused on machine learning and software engineering.
Additional coursework completed in computer science.
While working full time, I engaged in Stanford's graduate-level computer science classes focused on machine learning, artificial intelligence, and data mining.
CodePath's intensive eight-week Android development bootcamp provided me with a solid foundation in Android development, allowing me to further enhance my software engineering expertise.
I led the development of innovative computer vision systems designed to enhance electric grid inspections, culminating in a patent for a novel system to improve grid management processes.
Signorotti, Michael, and Maryam Variani. "Generative AI for Predictive Maintenance." Generative AI Community Conversation, Utility Analytics, 9 Jan. 2025. Conference Presentation.
Signorotti, Michael, and Kunal Datta. "Artificial Intelligence Enabled Aerial Inspections at PG&E." Utility Analytics Summit Reimagined, 9 June 2020. Conference Presentation.
Signorotti, Michael. “Feature Engineering for Time-Series Electric Interval Usage Data.” Association of Edison Illuminating Companies. Load Research and Analytics Workshop, 20 Mar. 2018, Lake Buena Vista, FL. Conference Presentation.
The Wall Street Journal featured my team's innovative work in leveraging computer vision to enhance PG&E's asset inspections. The article details how our models help improve efficiency and reduce wildfire risks, with insights from two of my colleagues.
As a member of the product development team for the Sherlock Suite, I contributed to a cutting-edge software system that combines a web application for utility asset review with a computer vision platform for enhanced inspection. My role focused on developing computer vision models to identify utility assets in high-resolution drone imagery and assess the health of critical hardware. This work earned PG&E the prestigious 2020 CIO 100 Award for IT Excellence.
I would love to hear from you! The best way to contact me about relevant opportunities is through LinkedIn.