Organizations drown in activity but starve for impact.
I build the teams, metrics, and systems that change that.
Leading LLMOps for Cisco's AI Org. Built Baker Hughes' first autonomous digital twin for drilling fluids. Tech lead for Optimal Digital Solutions at GE. Stanford PhD.
Experience
Cisco
Leading LLMOps for AI Canvas and AI Assistants — developing and executing strategy for evals, A/B testing, and agentic workflows.
Defined and executed end-to-end data & AI strategy for Cisco's global manufacturing supply chain, supporting Cisco's global product revenue. Led a cross-functional team of 10+ across data science, ML engineering, and analytics. Built and delivered automated inventory-forecasting, optimization-based PCBA allocation, and supply-chain GPT.
Baker Hughes
Led a team of 5 engineers to create Baker Hughes' first autonomous fluid-management digital twin, integrating ML, IoT, and real-time optimization. Secured external co-funding and delivered measurable per-well profitability gains. Published 7 peer-reviewed papers and filed 4 patents on digital-twin and AI-driven drilling technologies.
GE — Optima Digital Solutions
Tech lead for cloud-native SaaS ML products that improved energy-services profitability significantly. Designed an AutoML framework (PyTorch + Bayesian optimization) cutting model-development cycles by half. Built digital-twin networks for BP offshore platforms.
QRI Group
Engineered data-assimilation and optimization toolkits that accelerated reservoir simulation by >100x, producing substantial ROI for clients.
Stanford University
Postdoctoral research in optimization and machine learning.
Built the optimization and ML foundations (5 journal papers, 700+ citations) later applied to enterprise-scale impact at GE, Baker Hughes, and Cisco.
Philosophy
The most important job of a leader is to set a clear vision and ruthlessly prioritize what matters. Without that, organizations default to high-volume activity that never moves the needle — teams ship point fixes and surface-level patches while the root causes go untouched.
I have seen this pattern repeatedly in large organizations: significant effort is spent, talented people are busy, yet key metrics barely move. The missing piece is almost never more activity — it is strong leadership that connects every initiative to the outcomes that actually matter.
This is especially true when building AI applications. Without a disciplined vision for evals, continuous feedback loops, and rapid experimentation, teams can iterate endlessly without improving the metrics that define success. The leaders who win are the ones who close the loop between customer feedback and developer focus — and create an environment where experiments are designed to move the key metrics, not just check boxes.
Impact
Tech lead for ML products at GE that drove measurable profitability gains in energy services. Designed AutoML frameworks that cut model-development cycles in half.
Defined and executed AI strategy for Cisco's global manufacturing supply chain across forecasting, optimization-based allocation, and GenAI-powered analytics.
Built automated inventory-forecasting systems that significantly outperformed legacy methods, enabling data-driven demand planning across global operations.
Led Baker Hughes' first autonomous fluid-management digital twin integrating ML, IoT, and real-time optimization. Secured co-funding and delivered measurable per-well profitability gains.
Filed 4 patents on digital-twin and AI-driven technologies at Baker Hughes, covering autonomous fluid management, prescriptive analytics, and real-time optimization systems.
Research cited over 700 times across optimization, ML, and computational science. Published 12+ peer-reviewed papers including a sole-author publication in the Journal of Computational Physics.
About
I lead AI/ML teams that deliver measurable business outcomes. At Cisco, I lead LLMOps for AI Canvas and AI Assistants — developing strategy for evals and A/B testing. Previously, I defined and executed the AI strategy for Cisco's global manufacturing supply chain, leading a cross-functional team of 10+ across data science, ML engineering, and analytics. At Baker Hughes, I built the company's first autonomous digital twin, secured external co-funding, and filed 4 patents. At GE, I was tech lead for ML products driving significant profitability gains and designed an AutoML framework that cut model-development cycles in half.
My Stanford PhD in the Department of Energy Science & Engineering gives me the technical depth to pair rigorous algorithmic thinking with pragmatic engineering — bridging the gap between research and production at scale.
Expertise
Enterprise AI roadmaps, cross-functional leadership, and taking prototypes from proof-of-concept to production at scale.
Building LLM-powered applications, AI assistants, and agentic workflows for enterprise deployment.
AutoML frameworks, model serving, monitoring, feedback loops, and production ML pipeline architecture.
Recruiting, scaling, and mentoring high-performing data science and ML engineering teams from inception to maturity.
Research
Shirangi, M. G. (2019). Journal of Computational Physics 390: 249-264. doi:10.1016/j.jcp.2019.04.003
Patents
Baker Hughes, 2019–2022. Covering autonomous fluid management, prescriptive analytics, and real-time optimization systems.
Contact
Interested in discussing AI strategy, optimization challenges, or collaboration? Feel free to reach out via email or connect on LinkedIn.