About
With a solid four-year background in machine learning and artificial intelligence, coupled with eight versatile years as a software development engineer across various tech environments, I am a fervent advocate for technology's power to shape society. I take pride in being a perpetual learner, a mentor, and a leader in the realm of data and computer science. My driving ambition is to contribute to the creation of intelligent, agile systems that integrate seamlessly into the social fabric.
Machine learning - AI - Product Engineering
My journey has been marked by a passion for deep-diving into tech from Machine learning, full stack, to IoT embedded system development, while embracing the philosophy of language-agnostic programming.
- Location: Boston, United States
- Email: yogesharjun94@gmail.com
A graduate degree in Artificial Intelligence opened me to the world of Machine learning and data science. With the software and product engineering experience coupled with the AI know how, I have a special inclination towards building machines/products that learn from themselves and making it part of the social cog – creating a difference.
Skills
Resume
Over the years, I have had interesting opportunities to work on. From working for a Casino gaming company to working on cutting edge research and data science at Harvard Business School, my career has been a whirlwind of technological experience. I have had a chance to do various full stack engineering, production deployments, data pipeline building, cloud ecosystem architecture, program management, mentorship and tech business consultancy. Here is my Professional Resume.
Professional Experience
Information Technology Scientist
Digital, Data and Design @ Harvard | Boston, MA
Jan 2023 - Present
- Cloud Architect. Designed, and engineered a data platform for the institute on AWS. This role involved detailed analysis and integration of cloud services, focusing on optimizing data management and storage with differential access, while ensuring scalability, security, and efficiency in performing data science from level of experimentation, research to real-time production.
- Data Engineer – helped team with data migration and Extract Transform and load operations of data from the legacy environment to the cloud. Built pipeline to move data of size few terabytes from on prem to query efficient data catalogue, NoSQL databases and buckets.
- Leadership in Process Optimization: Spearheaded the enhancement of source control management and deployment tactics, successfully boosting research throughput and establishing a framework for the modular and repeatable application of data science solutions. This initiative not only streamlined operations but also significantly improved the efficiency and consistency of project outcomes.
- Rapid Development and Deployment: Successfully engineered and deployed a cutting-edge Retrieval Augmented Generation (RAG) based Generative AI Chat application within an impressive timeframe of just 10 days. This project involved building the application from scratch, ensuring its seamless integration into production environments, and demonstrating a high level of technical proficiency and efficiency in software development.
- Agile AI Development: Developed a Generative AI based simulation script capable of distributed computing, designed for idea generation and prompt evaluation. This advanced tool integrates multiple large language models, including GPT-4, Claude, and Llama 2, showcasing my expertise in handling complex AI algorithms and optimizing them for high-performance distributed environments. This lead to the discoveries on the Cyborgs vs Centaurs paper.
- Innovation in Language model: Lead a team in pioneering a unique self-instruction method for large language model fine-tuning, integrating RAG at the foundational level and PEFT-based QLoRA (Prompt Engineering and Fine-Tuning lower ranks) at the higher end. This groundbreaking approach culminated in the development of a white paper, code-named CRISP-LLM. This paper presents a framework designed to construct an open-source, domain-enriched language model, marking a significant contribution to the field of AI language processing. Process including workflows to domain selection, creation of fine tuning data with RAG, LLM and human in the loop, finetuning, validation with inter-rater reliability and LLM benchmarking.
- Machine Learning Architect: Devised and developed a sophisticated, cloud-based Generative AI application on AWS, featuring a containerized web application. This solution seamlessly integrates a serverless large language model inference engine, leveraged by both vector and NoSQL databases, ensuring efficient data handling and scalability. Additionally, the architecture includes scalable modular server components, all tied together with a robust CI/CD pipeline for streamlined development and deployment.
- MLOps: Designed and implemented a comprehensive set of data science cloud development practices across the organization, significantly enhancing the research efficiency of non-data science researchers. This initiative was met with exceptional success, achieving a 100% adoption and acceptance rate. By streamlining workflows and providing user-friendly tools, the project not only democratized data science capabilities but also fostered a culture of innovation and collaboration across various departments.
- Technology Mentor for the Startup Observatory, consulting with budding entrepreneurs on Tech Stack and engineering. Helping with set up for success with 0% attrition and increasing client growth by 2x.
Research Assistant and Program Manager
IoT Open innovation labs @Northeastern University | Boston, MA
July 2021 - December 2022
- Project Leadership: As the Project Lead, I spearheaded the development of a dynamic job marketplace backend using FAST API and SQLAlchemy. My responsibilities included the construction and orchestration of Docker containers on AWS, efficiently integrated with a CI/CD pipeline utilizing GitHub Actions. Additionally, I designed and delivered an engaging and user-friendly frontend using Vue.js and Vuetify.
- Architected a cloud solution and engineered an analytics engine for the platform with unstructured data store and live data pipeline. The engine would consume clickstream and user data on multiple fronts (custom designed, google analytics and Microsoft clarity).
- Innovative Research Platform Development: Played a pivotal role in defining and creating an A/B test embedded research experiment platform. My contribution encompassed both the system design and its implementation within an ambitious timeframe. This work involved developing a robust and scalable system architecture tailored to facilitate efficient and accurate experimental testing, demonstrating my proficiency in rapid, yet meticulous, project execution.
- Program Management: Led a team of four, skillfully crafting a work breakdown structure to optimize task allocation and efficiency. Established clear agile milestones and a strategic roadmap to guide project progression. Additionally, I implemented an event-driven monitoring solution tailored for the scrum team, enhancing our ability to track progress, anticipate challenges, and adapt swiftly to changes. This approach not only streamlined project management but also fostered a collaborative and responsive team environment.
Software Development Engineer Coop
Amazon Robotics | Boston, MA
January 2022 - June 2022
- Automation and Integration Expertise: Developed an advanced script designed to automate robotic operation simulations, significantly enhancing the accessibility of existing services. This script provided seamless programmatic access, facilitating easier and more efficient user interactions with our service offerings. The implementation of this automation not only streamlined operational processes but also played a crucial role in elevating the overall user experience, demonstrating a strong commitment to technological innovation and user-centric design.
- Developed a statistically sound machine learning optimizer based on the finite difference method for gradient descent. This tool was specifically engineered to generate optimal parameters and configurations for designing warehouse floor plans. Its implementation resulted in a substantial reduction in planning costs, showcasing the practical application of advanced machine learning techniques to streamline complex logistical challenges and enhance operational efficiency.
- Skillfully fabricated a robust infrastructure-as-code framework to meticulously engineer a sidecar service container operating within AWS Batch. This initiative involved breaking down building blocks of containerization to redesign a highly efficient, automated environment for the existing containerized service. The successful deployment of this infrastructure demonstrates my proficiency in cloud-native technologies and my commitment to creating resilient, and cost-effective cloud solutions.
Senior Software Engineer
Kloudone | Chennai, India
October 2019 - December 2020
- As the Team Lead for a group of 10 engineers, I effectively managed team operations, resource allocation, and authored user stories in an agile environment. Additionally, as a Training Lead for 30 engineering trainees, I developed comprehensive training modules and established coding best practices. These guidelines became integral to application development across the organization, reflecting my ability to mentor and guide large teams in complex technical environments.
- Full-Stack Application Development: Successfully implemented a SASS application utilizing a diverse tech stack including React.js, Gatsby, and Flutter for iOS and Android. This project showcased my versatility and expertise in full-stack development, cross platform development and analytics integration.
- Product Engineer: Led a product design sprint, where I was responsible for defining the Product Requirement Document and framing the Product Roadmap. This strategic approach facilitated clear visioning and planning, ensuring the product's alignment with market needs and organizational goals.
- Cloud Ops and Observability: Orchestrated the infrastructure setup on Google Cloud Platform, managing server runtime, SSL certificates for wildcard and sub-domains, and application deployment. Additionally, I implemented Application Performance Monitoring using the ELK stack.
Software Engineer
Freelancer | Chennai, India
December 2017 - October 2019
- As a consultant, liaised with external stakeholders to define scope and design roadmap. Constructed a website and a companion mobile app. UI/UX design on Figma, frontend on Angular, server and backend on LAMP stack with companion app on Flutter. Programmed schedule management service with PHP.
- Managed a team of 5 engineers and Architected a MEAN stack application with Angular 8 on Typescript. Application was deployed as a Progressive Web Application with Server-side rendering enabled using Universal framework. State management done using ngStore and RxJS. Backend was built using Express.js and Sail.js running on a node,.js environment to run as a side car.
Associate Software Engineer
Scientific Games | Chennai, India
June 2016 - November 2017
- Structured a software that was deployed as a desktop application using open-source Chromium engine.
- Forged a backend that was engineered with .Net ASP MVC, MS SQL Server, and ISS.
- Collaborated to web frontend that entailed working on .Net Razor, Angular 1.x, d3.js and jQuery. Built an animation library using CSS3 and Document Object Model. Built a Data Table template using Angular for cross organizational consumption.
Certifications & Publication
TATA Consultancy Service Best Project
2016
On device IoT smart assistant using Natural language processing and Smart Fidelity with Speech Recognition on a Raspberry Pi.
Big Data Analytics using Hadoop Ecosystem
2015
Conducted by Centre for Development for Advanced Computing (CDAC), a Scientific Society of the Department of Information Technology, Government of India. Graded A in Certification examination.
Health Monitoring Systems by Prognotive Computing using Big Data Analytics
2015
A computing system that does prognosis with machine learning and soft computing techniques. Published in Procedia Computer Science.