Victor Gueorguiev

I'm excited by the opportunity to join SAP's Hyperspace Product Management Solutions team as a Senior Machine Learning Engineer, and to help build the generative AI and multi-agent systems that will support product managers across SAP Engineering. I believe my hands-on experience designing, building, and deploying production machine learning and agentic solutions, combined with my experience in software engineering, in MLOps, and in leading a team, makes me a great fit for what you are looking for in this role.

In my most recent role as a senior machine learning engineer at Living Homes, I have been responsible for the product's AI capabilities in the cognitive backend layer, owning end-to-end feature development on multiple critical-path features and working directly with product owners from user story and design through to QA handover. Among the solutions I have built and shipped to production are a constrained conversational smart-home LLM agent with tool and function calling to internal microservices and external APIs, reinforcement-learning-based sleep-environment optimization, statistics-driven sleep and health analysis, AI-based user voice identification, and a dynamic home-automation rule engine. I believe that this maps closely to what the role needs: designing and deploying scalable Gen AI and multi-agent systems on a cloud and MLOps stack (Python, Azure, Docker, Kubernetes, Temporal, and LangChain), with reliability and observability in mind.

Prior to that, as a senior data scientist and team lead at VMware Carbon Black (now the Enterprise Security Group at Broadcom), I was responsible for the design, implementation, deployment, project planning, and stakeholder communication of various analytical solutions and predictive models in the space of customer analytics, KPI reporting, process improvement and product cost optimization, and in novel algorithm design and proof-of-concept generation. Some noteworthy projects to mention at a high level included building analytical metrics measuring customer adoption KPIs, building KPIs for measuring engineering support efficacy in resolving customer-reported product defects, building and serving a customer churn model, creating product cost anomaly monitoring tools, building EDR traffic monitoring tools to reduce product costs by 10-15% (to the order of millions of dollars), researching use-cases for graph-based threat detection, and researching large language models (LLMs) as a technology to create a cybersecurity virtual assistant and an internal self-serve BI tool for analytics users.

In my time at Living Homes, I led feature development end-to-end, and implemented them using Azure as the cloud platform, Terraform for IaC, and worked with frameworks such as FastAPI, Flask, MongoDB, Redis, Temporal, Langchain, MLflow, and more, in order to build our features in a microservices setting. At VMware Carbon Black, I served as technical lead in designing the technical solutions and their subsequent implementation and architecture, and as team lead for the team implementing them. For their implementation, we used AWS extensively, using services in AWS such as Glue (Spark), Kinesis, Athena and SageMaker for the backbone of our solutions. After VMware was acquired, we began a large migration workstream to GCP, using a mirror stack of services to migrate our data solutions.

Across my professional experience in product companies, and in my years in consulting before that, I have worked closely with product owners and I have been exposed to a wide range of project stakeholders, and I have subsequently picked up the necessary skills to manage the intricacies that come from having diverse stakeholder profiles and vested interests in the project outcomes. As a data scientist in consulting, I learnt that effective communication of the team's work to non-technical stakeholders, and effective interpretation and scoping of project requirements from the same stakeholders, is critical. And taking that to the next level, working in product companies, I needed to bring together engineering teams, product owners, and other vital stakeholders to ensure that the feature being delivered is up to spec, or that the correct KPIs were being tracked, and overall that the best value was being delivered via our features/projects.

I believe my diverse experience in machine learning and generative AI across different problem domains, and my in-depth knowledge of modern machine learning deployment stacks and cloud infrastructure, lends me to be an excellent candidate choice for SAP, and I trust that my skillset will be a unique addition to the team's portfolio.