Siva Sravana Kumar Neeli
Profile

I’m a Machine Learning Engineer with expertis ranging from experimentation/insights to Deep Learning frameworks transforming the latest research into production-ready solutions. My OSS contributions stand out by empowering a global developer community and driving next-generation AI innovation. My contributions : https://github.com/sineeli

Skills
Programming & Data Manipulation: — Python, C++, Java, JavaScript, HTML, ML Frameworks — Keras, Tensorflow, JAX, Torch, HuggingFace, LLM's, Tools & IDE — VSCode, Github, MLflow, Gitlab, Git, JIRA, API Frameworks — FastAPI, Django, Cloud — GCP, AWS (familiarity), Vertex AI, Cloud Run, Kubernetes (basic), Databases — PostgreSQL, SQLite, PGVector, SQL (BigQuery)
Professional Experience
Machine Learning Engineer, Tech Mahindra Americas, Google(Contractor)
August 2024 – present | San Jose, CA
  • •Implemented Faster R-CNN, RetinaNet and Vision Transformer(ViT) in KerasHub using Keras3, enabling seamless backend switching between JAX, Torch and Tensorflow.
  • •Developed a Colab notebook leveraging Gemini API(LLM) to automate generation of MCQ and short-answer-questions from pregnancy care guidelines, showcasing AI-driven survey automation for improved data collection and analysis.
  • •Constantly analyze user feedback on AI Studio and the Gemini(LLM) API SDK, identifying gaps and opportunities to enhance the product. I relay insights to the product team and contribute ideas to align with customer needs, driving better adoption and user satisfaction.
  • •Contributed 35+ Pull Requests across Keras, TensorFlow and KerasHub enhancing model compatibility and performance.
  • •Presented weekly pain points from the user community to the respective product teams, prioritizing and resolving potential future issues, while leveraging strong presentation skills to clearly communicate user needs and influence product decisions.
    • Machine Learning Engineer, Tana Systems Inc, Google(Contractor)
    June 2022 – August 2024 | San Jose, CA
  • •Resolved repetitive GitHub issue troubleshooting for a 40-member team by building an automated recommendation system that identifies similar closed issues, leveraging PGvector VectorDB for scalable updates, and cutting resolution effort by 30%.
  • •Developed a scalable Mask R-CNN & YOLOv8 pipeline on Google Vertex AI with distributed data-parallel training (scales up to 8-GPU's) on a 175GB dataset.
  • •Authored 5 advanced tutorials on TensorFlow Model Garden, streamlining distributed training and data augmentation to accelerate Google’s AI development and fostering OSS.
  • •Led code reviews for 2 team members, ensuring coding standards, and pioneered new ideas/documentation for Google’s OSS (TensorFlow/Keras) to benefit users.
    • Data Scientist Module Lead, GlobalData Plc
    August 2021 – May 2022 | Hyderabad, India
  • •Developed predictive models (AR, MA, ARIMA, SARIMA) to forecast coal & steel production for the next 5 years with 70% accuracy.
  • •Enhanced client reports with data-driven forecasts, providing deeper market insights and improving user satisfaction.
  • •Developed a Django-based active learning POC for labeling massive raw dataset, leveraging iterative SME feedback to continually refine base model accuracy.
    • Machine Learning Engineer, Brane Enterprises Pvt Ltd
    July 2020 – August 2021 | Hyderabad, India
  • •Led a large-scale migration from Django to FastAPI with OOP for more code reusability, which led to 20% less lines of code.
  • •Developed a high-volume semantic search pipeline for BERT/RoBERTa models with FAISS indexing, enabling near-instant retrieval.
  • •Migrated the entire codebase to GitLab, creating automated CI/CD pipelines (Docker builds, testing) with YAML config.
  • •Led a four-member team, mentoring on coding best practices and version control to enhance productivity and project delivery.
    • Image Processing Engineer, Statvalu Pvt Ltd
    December 2019 – July 2020 | Bengaluru, India
  • •Engineered a CNN model for identifying question-answer areas within scanned form insurance images and classify them into Yes/No.
  • •Created a module using OpenCV for precisely locating small checkboxes and radio buttons in low-resolution scanned forms, achieving an impressive F1 score of 0.96.
  • •Enhanced the accuracy of OCR output from damaged documents by employing a range of advanced image processing techniques.
    • Assistant System Engineer, Tata Consultancy Services(TCS)
    July 2016 – October 2017 | Mumbai, India
  • •Enhanced and maintained banking systems by adding new features and fixing issues in the COBOL codebase.
    • Education
    M.Tech(Data Science), Dhirubhai Ambani Institute of Information and Communication Technology(DA-IICT)
    2018 – 2020 | Gandhinagar, India
    B.Tech(Electrical and Electronics Engineering), V.R.Siddhartha Engineering College
    2012 – 2016 | Vijayawada, India