Smart manufacturing & Industry 4.0
Active learning / adaptive experimental design
Bayesian optimization & uncertainty quantification
Data-driven process mapping for complex unit operations
Interpretable ML (symbolic regression, physics-informed models)
Digital twins and real-time process control
Laser powder bed fusion and other additive-manufacturing processes
Process systems engineering & optimization
Industrial IoT integration and edge analytics
Industry 4.0 demands engineers who can fuse rigorous process knowledge with modern data science. My research blends first principles engineering insight with modern Bayesian and information-theoretic learning to build accurate process maps when experiments or simulations are expensive. Our workflow runs entropy-guided batch active sampling to gather the most informative data, then applies symbolic regression at the end to distill closed-form, physically meaningful models.
We use a pulse-wave laser powder bed fusion (L-PBF) process as the experimental testbed. Starting from only a handful of single scan track (SST) trials, the active learner rapidly focuses on the transition mode region – where part density and variability are both high – without the combinatorial overhead of full-factorial or dense grid designs. The resulting symbolic models expose closed-form relationships among laser power, scan speed, laser pulse frequency, and melt pool morphology, making the findings suitable for real-time control and digital twin deployment.
The workflow is process-agnostic. I have applied a similar strategy to simulated phase diagram mapping for multicomponent systems, again showing that adaptive selection can delineate critical boundaries with far fewer evaluations than naive sampling. Because it couples statistics, optimization, and first principles reasoning, the approach scales to any continuous or batch operation where time- or cost-intensive testing limits development.
I am eager to translate this blend of experimental design, Bayesian optimization, and robust Python workflows into industrial roles focused on smart manufacturing, digital twins, and broader process systems engineering challenges. My goal: turn data-lean operations into data-driven ones – safely, efficiently, and at scale.