Cloud Architect
2020 - PRESENT
Responsible for cloud migration and modernization of products serving the Energy Market Porfolio which includes the following industry leading products:
- PROMOD Simulation Engine,
- Velocity Suite Energy Market Intelligence,
- TRMTracker Trading and Risk Management,
- nMarket Market Operations,
- Retail Operations,
- Counterparty Settlement and Billing.
Responsible for the Cloud Architecure with focus on cloud cost efficiency, performance, scalability and developer productivity.
Expertise in Google Cloud platform, Google Cloud Foundation Blueprint, BigQuery, Data Pipelines, Agigee API Gateways, Cloud Build, and Cloud Run.
Focus on Infrastructure as Code.
2007 - 2019 and 2001 - 2004
As a research scientist my role at ABB is to design, build, evaluate product features and concepts to benefit various ABB Business Units Industrial Automation: Oil & Gas, Pulp & Paper, Power Generation; Power Systems: Electric Grid Network Management, Micro-Grids; Robotics.
Work on strategy, product features, future vision with Product Managers, Technology Managers, Research Managers, & ABB Customers.
Projects ranged from using Machine Learning, Embedded Systems, Software Architecture, Software Performance, InfoSec, Full Stack Development, Mobile & Cloud Computing, Scalable Systems to improve & enhance software capabilities at ABB.
Published papers, invention disclosures & patents for advances in Industrial Automation & Analytics. Designed, implemented & presented various prototypes at ABB Customer World events to ABB’s customers.
Lead editor for a quarterly report on Software Technology Watch to evangelize Cloud Computing, Machine Learning, Operating Systems & Programming Language advancements to the larger ABB community.
1998 - 2020
Led development of three major updates of website for the largest Islamic center in South East with peak visits of over 10,000.
Linux, MySQL, NGINX, PHP, JavaScript, JQuery, MovableType technology stack. Built various REST endpoints for calendar & prayer services. Exploring Flutter, Jekyll, Hugo, Firebase for the next evolution of the site.
Built GSuite based IT infrastructure for over 100 users for day to day operations of the center.
2005 - 2006 and 1995 - 1999
As a senior architect in IBM Global Services, I worked on two major projects at IBM customer sites, for modernizing workflows for insurance claims & point of sale terminals. Built integrations with various IBM legacy systems using IBM MQ Series, IBM Websphere, & Guidewire.
In my previous stint at IBM as a Senior Developer, I was part of the team that built VisualAge family of IDEs for Smalltalk, C++, Java which later became Eclipse IDE. Built visual composition editors, wizards & code generators using Smalltalk & Java.
2000 - 2002
Founded a technology consulting company providing software development & training in Java EE + Web development. Built developer certification exams for IBM.
Customers included IBM, ABB, Vroom Technologies, Local & State Government Agencies, and news services. Built & delivered tutorials & course work for Java EE, EJB, Struts, JSP, Servlets, XML, UML, OO Design.
1989 - 1995
Taught VAX 8600 Assembly Language & Data Structures classes to undergraduate students. Mentored students with internship & job placements at RWD Technologies.
1992 - 1995
Built custom software solutions for Holiday Inn, Motorola, GE for Macintosh, Windows & UNIX platforms. Wrote custom widgets & built a cross-platform user interface & communication APIs for GEnie online service. Built custom parsers for domain specific languages.
1990 - 1992
Built nuclear power plant simulators. Focused on Instructor Workstations to override simulations to introduce faults. Used real-time UNIX on Masscomp workstations. Ported Simulator Instructor Workbench to Sun Solaris OS. Wrote a DataViews overlay to help automate labeling engineering drawings.
1989 - 1990
Master’s thesis on implementing & measuring performance of various graph algorithms. Presentations on dealing with Byzantium Faults in Distributed Systems, Public Key Cryptography & Zero Knowledge Proofs.
1984 - 1988
Final year project was Universal Assembler design & development in Turbo Pascal. Taught Pascal & Z80 Assembly Language programming in senior year, this was my very first job. Teaching is my passion.
- Cloud Computing Architect, Google Cloud, OpenStack, Docker, Kubernetes.
- Full Stack Development & Mobile App Development using Flutter.
- Linux Embedded Systems & Network Programming.
- Software Defined Networking, Network Func. Virtualization.
- Applied Machine Learning, Google Cloud APIs.
- Software Performance, Microbenchmarks, Perf tools.
- Teaching, Public Speaking, Writing.
- Software Technology Watch, Fuchsia OS, Swift LLVM, TensorFlow, QEMU.
- Compilers, Sanitizers, Fuzzers, Build Systems.
- Promoted to Senior Principal Scientist, second highest position at ABB Research.
- Awarded four US Patents in the field of Industrial Automation & Asset Analytics.
- Received Governor's Award for ranking first among all university students.
Go, C, C++, C#, JavaScript, Java, Dart, Python, Swift.
- Vim & Sublime Edit.
- Bash & Fish.
- Bazel & GN.
- Clang & LLVM.
- Git Command Line.
- Go Language Ecosystem.
- Google Cloud Platform & GSuite.
- Modern C++ & Abseil.
- SQLite, Dockerfile, Nginx Proxy, GraphQL.
- Linux Filesystem & Linux Networking.
- Chromium Mojo & Fuchsia FIDL for IPC.
- Scoop.sh on Windows & Brew on MacOS.
- Dart & Swift & Flutter Mobile Apps.
- Reading, understanding, and writing code.
- Chrome Developer Tools.
Software Defined Networking SDN Switch — Dept of Energy, Microgrid Resiliency Project at ABB Corporate Research
Built a physical SDN switch using Linux, Intel 4-port NIC PCIe Card, Linux Kernel Modules, Hugepages, VFIO-PCI, Open vSwitch, Intel DPDK, Faucet SDN, Prometheus & Grafana.
The SDN switch allows us to detect malfunctioning microgrids and isolate them using distributed Power Grid State Estimation algorithms.
This project helped me understand low level Linux Kernel Modules, Open Flow, Networking Namespaces, VLANs, TUN devices, Bridges, Gateways & Routers. Experimented with some off the shelf SDN switches, P4 Language & P4 Runtime. https://micro-grid.github.io/swtch/
Explored the solution space for possible intermediate representations of customer provided inputs for ML based pipeline which included Swift, CUE, Skylark. Used Git-LFS to store & manage PDFs & other artifacts. Used Azure DevOps, GitLab, Go, GraphQL.
Built word embedding ML pipeline for classifying ABB spare parts, part items, location inventory. Automated labeling for over 1.5 million items. Improved spare parts fingerprint processes for all of ABB’s products. Used Go, Python, Jupyter Notebook, Keras, GitLab, Portainer, Webpack, Babel, ESLint, SQLite.
Applied Voice Assistant APIs for embedded field devices such as Aquamaster Water meter for better interactions with field operators. Raspberry PI plus Google AIY Kit connected to water meter using Modbus protocol over serial port. Used C, Python & Google AI APIs.
Presentation on how to build a machine learning pipeline, when should you use ML, how to prototype it, how to be aware of biases. Source material from Google’s PAIR project, customized for ABB.
Company wide introspection on the future challenges posed by AI, Deep Learning, & automation. Part of a global team looking at identifying potential use cases for ML based pipelines. Reviewed Google AI, Jeff Dean, Google PAIR, Microsoft AI research papers and talks.
An approach to migrate legacy monoliths to microservices, when, why, and how. Built prototypes & proof of concepts of various approaches & design patterns using Docker & Kubernetes. Focus on scalability and discovery of services. Comparisons between Azure & Google Cloud.
Inter-process communication is key to various products deployed on small footprint devices. Identified different scenarios & built micro-benchmarks using Google Benchmark on Raspberry Pi & Intel NUC platforms.
Results of the study were very surprising, Google Protocol Buffers & Google FlatBuffers were the winners over other IPC serialization techniques.
Investigated using perf tools and eBPF to find out why performance stalls, the unsurprising answer was L1 Cache misses.
This project utilized Bazel, C, C++, GTest, GBenchmark, perf & other Linux toolsets, it was the most appreciated project at ABB Research.
Built a mobile version of the ABB 800xA Control system to help field operations. Project involved using Go middleware, Material Design Progressive Web App. Go server used a ngrok reverse proxy endpoint to a privately hosted 800xA industrial automation control system.
To improve field operations, we deployed Bluetooth BLE Beacons throughout a factory floor to passively identify the control system assets closest to the field operator.
Investigated & applied for a secure geo fenced lifespan patent for a mobile device to be used in the field for 800xA control system operations.
Built a mockup prototype of power generation, water distribution autonomous asset inspection based on ideas from Google Street View and various image recognition techniques.
Built a modern progressive web app to highlight hotspots, indoor maps, onboard camera views, & potential diagnosis. The autonomous system relied on human operators to verify results & generate work orders.
Led a team of domain experts and research scientists to evaluate potential use cases for future mobility workforce. At this time, ABB was focused on desktop & panoramic control systems. Cloud & Mobile Computing were not considered essential.
We interviewed ABB’s customers in Oil & Gas, Power Generation, Power Distribution, Water Distribution, Industrial Automation to identify gaps and needs for field operations.
Web logs present a unique perspective on how web apps and websites are being used. We used these ideas and control room chatter to annotate and tag operator actions to learn which parts of the system are being used, which are rarely used, and other leading indicators.
Designed, prototyped, and implemented a new controller hardware & software platform for Industrial Automation. This was before the advent of mobile platforms and Linux on ARM platform.
Technology stack consists of IBM Java J9 Virtual Machine, Apache Jetty, IBM MQTT. Product rollout failed due to SD write performance issues and hardware reliability.
Stora Enso is a large paper & pulp factory based in Sweden. Designed and implemented a work order management system using Skyva plus IBM SanFrancisco framework (precursor to IBM WebSphere), using Java servlets, JSP, and other related components.
ABB was focused on Industrial IT and SAP competitor Skyva for building the next generation Manufacturing platform. Built Java middleware based on Skyva, SAP connectors, 800xA Control System integrations.
Built a static website generator using Java for Middle East & North Africa news content. Precursor to Jekyll and Hugo. Used templates and a simple tags to allow news editors to write content in text files and generate static website.
- Microinteractions in Mobile and Wearable Computing, published by IEEE Advanced Information Networking and Applications in 2016.
- Location Aware Mobile Systems, published by ABB Research USCRC Technical Review in 2015.
- A dynamic software product line architecture for prepackaged expert analytics, published by WICSA in 2014.
- Picture of health, published by ABB Review in January 2014.
- Exploiting ABB knowledge on industrial assets, published by ABB Review in 2013.
- Machine Learning at ABB Customer World, Houston Texas, March 2019.
- Human Centered ML: An approach to Machine Learning projects, ABB, 2018.
- Mobile 800xA Demo at ABB Customer World, Houston Texas, March 2017.
- Fuchsia at ABB, Capability based operating system, 2017.
- Flutter for Mobile Apps, Building native iOS, Android, Fuchsia apps from a single source codebase, 2017.
- Docker Containers Seminar, Using containers to build scalable systems, 2015.
- Go Slices and Bytes, Talk at Google Developers Group Triangle in Durham NC 2014.
- Using Struts at Java Users Group in Durham NC.
- Java EE, XML, UML technology courses at IBM sites throughout 2000 - 2001.
- Model development environment for assisting user in developing model describing condition of industrial asset, August 16, 2016.
- Industrial Asset Health Profile, December 4, 2014.
- Industrial asset event chronology, March 13, 2013
- Interactive interface for asset health management, June 25, 2015.
- Industrial asset health model update, December 11, 2014.
- Recording and providing for display images of events associated with power equipment, November 20, 2014.
- Criticality profile for industrial asset, June 10, 2013.
- Installing Arch Linux.
- Enrolled in Windows Insider Fast Ring, MacOS Beta, & Android Beta.
- Building my own Chromium browser on Ubuntu.
- Exploring using Coral TPU for Machine Learning on the edge.
- Built apps for Google Glass.
- Built super upgraded desktop with AMD Ryzen 16 cores, 128 GB DRAM, 1 TB NVMe, 5 TB SSD, AMD Radeon Vega GPU. Ideal for running 100s of Chrome tabs, TensorFlow ML, building Chrome & Fuchsia.
- Ask me about Spinlocks, FUTEX, File Locking, Go Scheduler, Network Namespaces, Containers, Fuchsia OS.
- Attended the Guy Steele talk at 1998 ACM OOPSLA Conference.
- Received a tape archive from Richard Stallman with GNU toolset in 1991.
- Using Go before its 1.0 release.
- Started working on web in 1998 with CGI scripts using Smalltalk.
- Started programming with C++ using cppfront preprocessor.
- Learned to program on a Z80 microprocessor kit using machine code.
You can reach me at shakeel at mahate dot net.