Valene Tjong Email & Phone Number

California, United States

Boston, Ma, Us

• Lead all ML projects as first/sole ML engineer with solutions powering opportunities for 6-figure contracts with some of the largest test prep/publisher companies in the world • Spearhead prompt engineering and multilingual, multi-modal model fine-tuning initiatives for a variety of complex AI solutions, executing end-to-end processes from design to deployment, including product, database management, A/B testing, and QA.• Architected, built, and set-up hosting for advanced multi-lingual RAG-application chatbot used by thousands of clients• Demonstrated ability to rapidly deliver optimized results independently, ensuring high-performance, client-ready implementations across all project phases.

Menlo Park, California, Us

• Writing research paper under publication.• Built robust ML (Gaussian Processes) infrastructure, adhering to MLOps best practices, to function as a recommender system for material fabrication processing conditions, automating the optimization process and achieving significant decrease in fabrication costs over previous pseudo-random sampling methods.• Conducted and automated NAS hyperparameter search on ML model, obtaining best model with minimum test loss.• Used multi-threading to schedule hyperparameter search process across CPU cores, reducing runtime by 7-fold.

Menlo Park, California, Us

• Conducted thorough and comprehensive literature review for optimum features to use for ML dataset.• Engineered and implemented modular data pipeline, used to efficiently cleanse, process, and visualize large-scale datasets for ML model, achieving 98% speed-up over previous Excel-based methods that entailed single feature-• Wrote ML portion of funding updates to DOE (Department of Energy).

New York, New York, Us

• Mentor and grade groups of undergrads in their respective projects • Create course material in building data analysis pipelines, training and validating ML models, etc.

Ithaca, Ny, Us

• Built and operationalized data pipeline and ML (LSTM, discrepancy modelling) architecture, using PyTorch, NumPy, Pandas, and OpenCV as well as conducted NAS (network architecture search) leveraging wandb automation to create digital twin, a neural network emulating the lab’s physical lasing system.• Trained digital twin using PAT (physics-aware training) to feed back to the physical laser, enabling setup to host other neural networks and complete machine learning tasks with computational efficiency of at least 103 times better than that of industry-leading digital supercomputers.• Modularized controls for all laboratory instruments (laser, arbitrary waveform generator (AWG), oscilloscope, and real-time spectrum analyzer) using Python wrappers and hosted them on lab-wide GitHub, as well as set-up server-client delivery, transforming data operation from analog control to digital, remote access, used by 50 researchers operating shared equipment (Code used in: https://arxiv.org/abs/2208.05088, refer to acknowledgements).

Satellite Beach, Florida, Us

+ Delegated weekly lesson plans and created practice materials for SAT I, AP physics, and AP Calc as well as college-level physics and math courses+ Brainstormed, outlined, and edited college application essays, successfully aiding > 15 students get into their dream schools

Ithaca, Ny, Us

• Created computational models in Maxima and Matlab to solve nonlinear differential equations, aiding in Prof. R. H. Rand’s research on the stability of DDE’s with various damping and time delay coefficients. • Developed numerical simulation model of population dynamics associated with different strategies in rock-paper-scissors, as a proof of concept for theoretical results.• Extended upon method of harmonic balance in order to develop novel way of approaching linear homotopies with time delay terms.

Master Of Engineering - Meng, Computer Science

Bachelor Of Science - Bs, Physics, Computer Science, Math (Minor)