About me (Registered since 25/02/2025)

I am a dedicated and results-driven professional with a strong passion for Artificial Intelligence (AI) and a deep commitment to advancing the field. As an aspiring AI specialist, with Master’s degree in Artificial Intelligence at Northeastern University. My educational background includes a Bachelor’s degree in Computer Engineering from the Milwaukee School of Engineering.

🔍 Technical Expertise:
I possess a diverse skill set, ranging from programming languages such as C++, C#, Java, Python, and more, to proficiency in machine learning, natural language processing, deep learning, and reinforcement learning. My experience extends to algorithm design, search optimization, and microcontrollers. I am well-versed in utilizing libraries like PyTorch, TensorFlow, and NumPy, and I am comfortable working on various operating systems and platforms.

🧰 Relevant Projects:
1. **Object Detection in the Automotive Domain (GitHub Link: [ObjectDetectionAutomotiveDomain](https://github.com/ashwinsharan158/ObjectDetectionAutomotiveDomain.git))**

2. **Scientific Paper Summarization and Simplification (GitHub Link: [Summarization](https://github.com/ashwinsharan158/Summarization.git))**

3. **Stock Prediction for Volatile Stocks (GitHub Link: [StockPrediction](https://github.com/ashwinsharan158/StockPrediction.git))**

4. **Reinforcement Learning Agents for Partially Observable Environment (GitHub Link: [SO-ISMCTS](https://github.com/ashwinsharan158/SO-ISMCTS.git))**

🎮 Work Experience:
As an AI Consultant at VisionForge, I projected $4.5M in revenue by enhancing custom OCR models and integrating LLMs. At Crowd Doing, I optimize data for LLMs, reducing response time and increasing output by 30%, improving model accuracy from 73% to 88%. At GE Healthcare, I enhanced AI imaging platforms, designed diagnostic tools, set up ETL pipelines, and automated testing, cutting costs by $9400 quarterly. My work spans AI, machine learning, NLP, and automation.

📖 Publication:
Techniques and challenges for summarization and simplification of scientific literature” is currently under review. (https://github.com/ashwinsharan158/Summarization/blob/main/paper.pdf).

I am eager to bring my expertise in AI, coupled with my strong analytical and problem-solving skills, to contribute to innovative AI projects and solutions. If you are seeking a dedicated AI professional with a proven track record in project development and a commitment to excellence, I would love to connect and explore potential collaborations. Feel free to reach out to me at sharan.a@northeastern.edu.

Skills

Tech Skills

Portfolio

Education

Key Skills and Competencies

AI, LLMs, Chatbots, ML, MLOps,

Work Experience

  • November 2024 - Present
    YWCA of Western Mass, United States

    Data Analyst

    Automated data workflows using Apache Airflow and SageMaker, saving 40+ hours of manual work. Cleaned and analyzed datasets to support model development and identify trends, enhancing customer support. Designed dashboards with Power BI to effectively communicate insights to stakeholders.

  • January 2024 - August 2024
    VizionForge, India

    MLOps Engineer

    I managed and optimized machine learning pipelines for intelligent document processing, utilizing tools like TVM and MLIR, and deployed classification and extraction models through Azure ML and Azure DevOps CI/CD pipelines. I integrated Azure Databricks and OpenAI NLP models to enhance document scalability and developed a corporate Retrieval-Augmented Generation (RAG) system using LangChain, creating a conversational AI for business summaries. By optimizing IDP solutions, I increased monthly revenue streams by $28K.

  • May 2019 - November 2020
    GE Healthcare, United States

    CT Reconstruction Software Engineer

    I maintained an AI platform for imaging software by enhancing the denoising deep learning platform’s speed through advanced parallelization with CUDA and OpenCL, modifying it for higher resolution using custom Gaussian filtering in the CNN, and implementing edge detection with a Sobel filter as a CUDA kernel. To improve the reliability of computer hardware, I designed diagnostic tools for testing network connections between subsystems, set up an ETL pipeline for field data analysis, integrated machine learning models into benchmarking processes, and deployed systems in Docker using C++, Python, and PyTorch. Additionally, I led a team to automate manual testing processes, saving $9,400 quarterly, and configured nightly tests on Jenkins via Docker using Python and Bash.

Languages

English
Professional
Hindi
Advance

Certifications & Licences