Senior Python Engineer - AI/Agentic Systems, Data Pipelines

Jerome Yutai Shen, Ph.D.

Sunnyvale, CA ***** Tel: 310-***-**** Email: *****.****@*****.***

LinkedIn: https://www.linkedin.com/in/jerome-s-986123113/ 14 years of experience in software development

Skills

Languages: Python, Golang(tooling/automation), Java, SQL, MATLAB, C/C++.

Data/Infra: Kafka, Flink, Spark, Hadoop, Azure (Data Factory, Synapse), AWS (SageMaker, EC2, ECS, RDS, S3)

Data Eng: ETL/ELT, data modeling, data quality rules, schema versioning, backfills, idempotent pipelines

Datastores: PostgreSQL, MySQL, Oracle, Redis, ClickHouse, BigQuery

Containers/CI: Docker, AWS ECR, Jenkins (Bitbucket Pipelines/Azure Pipelines)

Observability/DevOps: DataDog, Sentry; structured logging, metrics, alerting Professional Experience

Johnson and Johnson, Santa Clara, CA via Real Staffing. 04/2024 – 06/2025 Software Engineer

● Modernized a legacy, script-heavy automation platform by refactoring Python 3.10+ around stable gRPC contracts, redesigning core components with clearer OOP boundaries, fixing concurrency/lifecycle correctness (async/thread cleanup, event-loop modernization, middleware binding drift). Impact: 65% less redundant code, 40% faster onboarding, 30% fewer incidents. Developed, deployed and operated the platform in Linux-based environments.

● Engineered a high-reliability Python orchestration layer for multi-node bring-up (nested SSH), state tracking, telemetry synchronization, and deterministic end-to-end verification; built Pytest/BDD harnesses producing expected-vs-observed evidence for regression and release qualification. Automated log session control on distributed test hosts via programmable APIs, replacing manual CLI workflows.

● Diagnosed hard-to-reproduce, non-crashing failures caused by interface/schema drift across services (gRPC/Protobuf + message schemas) that dropped actions without errors, implemented schema-version checks (Pydantic) and CI contract validation gates to prevent regressions.

● Delivered run_id-based traceability + observability (run_id, versions, config, hardware, artifacts), enabling reproducibility, trend analysis, and faster cross-team debug loops.

● Shipped two data pipelines: (1) Django/PostgreSQL (AWS RDS) ETL that canonicalized metadata + artifacts across 1,000+ verification runs (PDF/Excel

structured schema via schema discovery), and (2) telemetry correlation that reconstructed per-run semantic events from aggregated logs to accelerate fault isolation.

● Containerized services and delivered CI/CD pipelines (Docker, AWS ECR/EC2, Jenkins, Bitbucket), standardizing build, release, and artifact management across teams.

● Developed Python-based instrument automation (drivers managers runner) with robust validation and fault handling; persisted run configuration and results in SQLite to enable reproducible execution and offline triage; added Pytest regression tests.

Stealth Startup. 07/2023 – 04/2024

Software Engineer

● Architected backend with Flask-RESTX (and prototyped Django REST services for structured ingestion), building an ETL pipeline that integrated OCR(extract, combining Azure AI Document Intelligence / Form Recognizer with Tesseract/pdfplumber for scanned PDFs), PII redaction/semantic normalization (transform), and Postgres/embeddings (load), improved document processing latency by ~40%, reduced manual review effort by ~60%.

● Built RAG pipeline (Azure Cognitive Search + OpenAI embeddings) enabling multi-turn QA with citation grounding.

● Automated deployments via Azure CI/CD; added Sentry-based error monitoring and structured logging, improving root-cause diagnostics.

Meta, Menlo Park, CA, via Pyramid Consulting Inc. 07/2022 – 01/2023 Software Engineer III

● Benchmarked 64-node GPU clusters (512 GPUs, 512 MPI processes) using NCCL collectives (AllReduce, Broadcast, ReduceScatter) over RDMA-capable RoCE fabric, achieving 90%+ bandwidth utilization and documenting scalability trade-offs across message sizes.

● Developed a Python log-processing pipeline to automate retrieval and parsing of large-scale distributed training logs (10–15 GB/run) from per-node stdout/stderr. Implemented async I/O with subprocess workers and regex extraction to produce schema-consistent pandas DataFrames, exported to JSON/CSV/XML for dashboards.

● Automated extraction of key performance metrics (bandwidth, throughput, latency) from parsed DataFrames, replacing ad-hoc shell scripts, reducing processing time from ~2 hours to <10 minutes, and improving accuracy from ~90% to 95%+.

Kaiser Permanente, Oakland, CA, via Pyramid Consulting Inc.10/2020 – 04/2022 Software Engineer

● Architected and built an interactive analytics & visualization platform for drug education coordinators at Kaiser Permanente, based on Python 3.7+

(Flask/Dash/Plotly), Redis, and heterogeneous data sources across Hive, Oracle SQL, PostgreSQL, and SQLite. The platform provided multi-tab dashboards with authentication/authorization, GUI-driven SQL query generation, data visualization, and integrated ML modules (e.g., matching analysis).

● Optimized read performance with Redis caching and improved reliability by resolving race conditions across worker nodes via Redis distributed locks.

● Integrated DataDog and Sentry for better error alerting and performance monitoring.

● Introduced Azure Front Door Content Delivery Network for rapidly delivering pictures and videos.

● Designed and extended ETL data pipelines and data flows in Azure Data Factory, generated tables in Azure Synapse.

● Leveraged Azure Pipelines to automatically build and test code projects for CI/CD.

● Ported the reported features to Power BI with a Direct Query option in SQL to get better insights on forecast.

● Leveraged Hadoop and Spark infrastructure on RedHat clusters to build large-scale ML workflows, using Spark MLlib alongside scikit-learn TensorFlow to improve patient cohort return-to-care predictions.

KLA-Tencor, Milpitas, CA, via Valiantica Inc. 12/2019 – 06/2020 Software Engineer

● Led a 5 people group to port the JMP apps to Python 3.7+, leveraged AWS SageMaker, pandas, numpy, scikit-learn, matplotlib and jupyterlab.

● Improved legacy codebase by constant refactoring, reorganization of dependencies, introduction of unit tests, integration tests, and end to end tests, achieved bug free and averaging twice faster than before, and high scalability with the framework Pytest.

● Developed APIs with Python Flask (existing services), FastAPI (new async endpoints) to provide data storage and access services, integrating AWS EC2 (compute), ECR (image registry), S3 (asset storage), and RDS (PostgreSQL metadata).

● Prototyped an ETL pipeline ingesting wafer equipment telemetry (Parquet) into ClickHouse, evaluating query performance for high-volume time-series aggregation and anomaly detection.

HP Labs, Palo Alto, CA, via Radiant Systems.10/2018 – 10/2019 Software Engineer

● Worked at Immersive Experience Lab which develops the next generation VR device, built the off-line and on-line analysis pipeline.

● Built a video transcoding system (which has six main components: preprocessor, DAG scheduler, resource manager, task workers, temporary storage, and encoded video as the output) to encode the videos collected from VR device into compatible bitrates and formats.

● Provisioned AWS EC2 instances with compute-optimized configurations to parallelize video transcoding tasks; utilized AWS S3 for resilient storage of intermediate chunks and final encoded assets, enabling asynchronous downstream processing.

● Implemented and optimized signal preprocessing and feature engineering programs employing zero-phase filtering, Hilbert transform, Welch method and multi-taper windowed FFT based Nonparametric Power Density Spectral estimation and leave-one out cross validation.

● Applied Continuous Wavelet Transform and Wigner-Ville transform to compute Spectrogram with PyWavelets and PyTFTB.

● Introduced CNN model and LSTM model to optimize feature engineering by dint of TensorFlow and Pytorch.

Shanxi Mount Heng Real Estate Development Limited, Taiyuan, China 06/2017 – 10/2018 Software Engineer

● Built an internal house recommendation and analysis system for the company, it can help our realtor to build better pricing and sale strategies.

● Created crawlers to collect descriptions and reviews of open or rentable houses from local realtor websites and structured the data to JSON.

● Calculated an inverted index table, leveraged Apache Pig to index on stemmed word and used Hive to load details table to HBase.

● Designed the DAG models which defines tasks in stages so they can be executed sequentially or parallelly.

● Implemented the crawler job scheduler to fetch houses’ new reviews and posted them into Kafka.

● Used Apache Storm to consume messages from Kafka and integrated the new data with existing records in HBase.

● Developed Golang scripts for cron jobs.

Department of Physics and Astronomy, UCLA. 11/2013 - 11/2016 Staff Research Associate 3

● Designed signal processing algorithms with MATLAB for filters design and zero-phase filtering, spectrogram with Continuous Wavelet Transform and Continuous Gabor Transform, instantaneous features with Hilbert Transform, denoising continuous signals.

● Applied dimension reduction machine learning algorithms such as PCA, t-SNE, and clustering such as k-means++, Mixture of Gaussians and Spectral Clustering.

● Adopted a Generalized Linear Model (GLM) to minimize the influence of behavioral biases, and obtained an unbiased estimate, with MATLAB.

● Ported a Python 2.7.x based standalone app to Python 3.3.x, deploying on OpenSUSE Linux to optimize MATLAB and Python scientific computing workloads.

David Geffen School of Medicine, UCLA. 11/2012 - 10/2013 Postdoctoral Scholar

● Developed experiment controlling and visual stimulus creation system in C/C++ 11 and RTOS QNX.

● Designed algorithms with MATLAB, combined spline interpolation, FIR filters designing and zeros-phase filtering for eye movement signal. Education Experience

Beijing Normal University, Ph.D. in Cognitive Neuroscience Beijing, China 09/2009 - 06/2012

Shanxi Medical University, M.Sc. in Physiology Taiyuan, China 09/2006 - 06/2009

East China Normal University, B.Sc. in Biotechnology Shanghai, China 09/2001 - 06/2005

Contact this candidate