𝗩𝗶𝘀𝘂𝗮𝗹 𝗜𝗻𝘀𝗽𝗲𝗰𝘁𝗶𝗼𝗻 𝗔𝗜𝙈𝙤𝙣𝙞𝙩𝙤𝙧𝙞𝙣𝙜/𝙏𝙚𝙨𝙩𝙞𝙣𝙜Ensure proper functionality and performance of pipelines► Created end-to-end API tests to verify functionality and performance.► Implemented automated bug reporting of probers falling below success threshold.𝙍𝙚𝙦𝙪𝙚𝙨𝙩 𝙑𝙖𝙡𝙞𝙙𝙖𝙩𝙞𝙤𝙣Certify requests are not only properly formatted, but also viable to run.► Authenticated billability immediately prior to training completion.► Blocked operations are that are unable to run concurrently.► Verified the customer has enough quota to execute the operation.

𝗔𝘂𝘁𝗼𝗠𝗟 𝗩𝗶𝘀𝗶𝗼𝗻𝘿𝙖𝙩𝙖 𝘾𝙚𝙣𝙩𝙚𝙧 𝙈𝙞𝙜𝙧𝙖𝙩𝙞𝙤𝙣/𝙍𝙚𝙜𝙞𝙤𝙣𝙖𝙡𝙞𝙯𝙖𝙩𝙞𝙤𝙣Allow batch prediction against customer trained models.► Seamlessly migrated and expanded to multiple data centers.► Ensured sufficient accelerator availability to scale with usage.► Updated pipeline configs to properly run with regionalization.► Built a quota tracking dashboard for real-time usage monitoring.𝘽𝙖𝙩𝙘𝙝 𝙋𝙧𝙚𝙙𝙞𝙘𝙩𝙞𝙤𝙣Allow batch prediction against customer trained models.► Benchmarked each model against various input sizes.► Integrated batch prediction and established tests.► Enabled the use of codegen to generate a model wrapper.𝙏𝙚𝙣𝙨𝙤𝙧𝙛𝙡𝙤𝙬 𝙇𝙞𝙩𝙚 𝙢𝙤𝙙𝙚𝙡 𝙚𝙭𝙥𝙤𝙧𝙩 𝙪𝙥𝙜𝙧𝙖𝙙𝙚Update TF Lite version and improve metadata.► Enabled the use of codegen to generate a model wrapper.𝙏𝙧𝙖𝙞𝙣𝙞𝙣𝙜 𝙋𝙞𝙥𝙚𝙡𝙞𝙣𝙚 𝙎𝙩𝙖𝙗𝙞𝙡𝙞𝙩𝙮Consistently increase the success rate of customer training.► Rewrote monitoring logic to include edge model training.► More accurately tracked actual user training time to prevent early termination.

𝗠𝗮𝗰𝗵𝗶𝗻𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗳𝗼𝗿 𝗖𝗼𝗺𝗽𝘂𝘁𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗖𝗵𝗲𝗺𝗶𝘀𝘁𝗿𝘆, 𝘎𝘰𝘰𝘨𝘭𝘦 𝘈𝘤𝘤𝘦𝘭𝘦𝘳𝘢𝘵𝘦𝘥 𝘚𝘤𝘪𝘦𝘯𝘤𝘦𝘴𝙎𝙢𝙖𝙡𝙡 𝙈𝙤𝙡𝙚𝙘𝙪𝙡𝙚𝙨 𝙐𝙣𝙞𝙫𝙚𝙧𝙨𝙚 𝙈𝙖𝙣𝙖𝙜𝙚𝙢𝙚𝙣𝙩 𝙎𝙮𝙨𝙩𝙚𝙢Versatile Management system to launch and monitor large-scale distributed computation tasks on a variety of cloud platforms.► Created, initialized, and tore down cloud machine instances as necessary► Minimized costs and time by utilizing preemptable VMs with automated retry► On average, this was able to reduce completion time by a factor of 6𝙈𝙪𝙡𝙩𝙞-𝙏𝙝𝙧𝙚𝙖𝙙𝙚𝙙 𝘾𝙤𝙣𝙩𝙞𝙣𝙪𝙤𝙪𝙨 𝙀𝙫𝙖𝙡𝙪𝙖𝙩𝙞𝙤𝙣Continuous evaluation workers running in parallel with training workers.► Established continuous distributed evaluation, removing step selection uncertainty► Provided more timely and detailed insight into overfitting of models.𝙏𝙚𝙣𝙨𝙤𝙧𝙛𝙡𝙤𝙬 𝙈𝙤𝙙𝙚𝙡 𝙎𝙚𝙧𝙫𝙚𝙧 & 𝘾𝙡𝙞𝙚𝙣𝙩Automated system to select the best trained Tensorflow SavedModels and launch a model server to run predictions against.► Allowed users to run predictions at higher QPS.

► Extended Command Line Interface to support additional commands► Maintained feature upgrades within internal fork of go-swagger and swagger-codegen► Refactored code and exposed RESTful API calls for public access

► Refactored code and exposed RESTful API calls for public access► Automated generation of API documentation using Go-Swagger► Implemented additional features in Go-Swagger for complete automation► Automated creation of API client code for testing and client use

► Manage SCCM clients for approximately 250 servers as well as a varying list of desktop machines.► Ensuring timely program and operating system updates for 200 servers with minimal downtime as well as verifying though QualysGuard that vulnerabilities are patched.► Monitoring, installing, and configuring Big Brother on a variety of unix-based servers.► Creating scripts and reports to expand the functionality of SCCM for day-to-day use.► Creating SCCM application packages for easy deployment of complex applications.

Unless otherwise stated, I was assigned to Alpha Company, 352d Combat Support Hospital.Typically worked 1 weekend/month and 2 weeks/year.► Learned and refreshed various soldier skills.► Led classes covering medical laboratory case studies.► Cross-trained medics to perform basic laboratory tasks.

Combat deployment with 349th Combat Support Hospital (Forward)► Managed laboratory services of five clinics and one hospital across every United States base in Kuwait.► Managed three Electronic Medical Records systems to keep patient data synchronized.

Combat deployment with 349th Combat Support Hospital (Forward)► Maintained operational status of all medical systems in Camp Buehring.► Transferred to Camp Arifjan to assist and restructure the hospital’s Information Technology department.

Advanced Individual Training Phase 2 - Clinicals

Advanced Individual Training Phase 1 - Academic

Basic Combat Training

Ranged anywhere from grading workbooks for students up to tutoring high school math and science