Aspiring ML Engineer

BHUVAN KARTHIK CHANNAGIRI

+1-857-***-**** Boston, MA **********.*@************.*** LinkedIn GitHub EDUCATION

Northeastern University, Boston, MA Dec 2025

Master of Science in Electrical and Computer Engineering Hardware and Software for Machine Intelligence Concentration Relevant Courses: Introduction to Machine Learning and Pattern Recognition, Natural Language Processing, Neural Networks and Deep Learning, Machine Learning Operations, Verifiable Machine Learning, Parallel Processing for Data Analytics R.M.K. Engineering College, Chennai, India May 2023 Bachelor of Engineering in Electronics and Communication Engineering Relevant Courses: Python Programming, Probability and Random Processes, Computer Architecture and Organization SKILLS

ML & Agentic AI: PyTorch, TensorFlow, LLMs, Computer Vision, NLP, Render, CrewAI, n8n, LangChain, RAG, ChromaDB, Flowise Data Engineering, Cloud & MLOps: GCP, AWS, Vertex AI, Render, Docker, Kubernetes, Apache Spark, PySpark, Apache Airflow, MLflow, DVC, Fast API, Git, CI/CD/CT/CM

Programming & Web Integration: Python, SQL, MongoDB, C++, C, React.js, Flask, HTML/CSS, MATLAB, Shell Scripting, REST APIs Data Visualization & Analytics: Tableau, Power BI, Grafana, Looker, Seaborn, Matplotlib, Plotly, Streamlit, Pandas, NumPy Bioinformatics & Image Analysis: CellPose, FIJI (ImageJ), Napari, Imaris, µSAM, Stata, DeepCell, Nextflow, BioPython, CyCIF WORK EXPERIENCE

ML Research Assistant, The Rowland Institute at Harvard, Cambridge, MA Jan – July 2025

• Engineered high-throughput image-analysis pipelines using ML and industry-standard tools (CellPose, FIJI/ImageJ, Imaris, µSAM for Napari), incorporating customized thresholding strategies to elevate segmentation accuracy to 90%

• Architected GCN models on large-scale graph datasets (50K+ nodes) to capture complex spatial relationships—expertise directly transferable to healthcare for modeling patient biomarker and cellular interaction networks, and proteomics data

• Designed graph-based GAN frameworks to optimize node-importance weighting and rigorously evaluate performance impacts

PROJECTS

Adversarial-Robust Vision Transformer for Histopathology Image Classification, [GitHub] - Boston, MA

• Crafted adversarial-attack modules (UAP, PGD, FGSM, Adversarial Patch) to rigorously probe model vulnerabilities and strengthen diagnostic reliability—quantifying attack success across 25 000 LC25000 histopathology images

• Established Vision Transformers as an adversarial defense, exploiting global self-attention to detect and mitigate perturbations—sustaining 92% of clean baseline accuracy under PGD, UAP, and adversarial-patch attacks and limiting performance degradation to under 8%

Ozone Level Detection (End to End Machine Learning Pipeline), [GitHub] - Boston, MA

• Developed and deployed a machine learning pipeline for ozone level detection on Google Cloud Platform, integrating services such as GCS, GCP, Airflow, and Vertex AI, resulting in a 30% improvement in deployment efficiency

• Implemented monitoring FastAPI, linking the model’s endpoint to GCP for real-time tracking and achieving 99.9% uptime

• Automated data processing and model training with Airflow DAGs, incorporating error handling and data quality checks, which decreased data processing time by 25% and improved model accuracy by 15%

• Containerized the model using Docker and optimized it through hyperparameter tuning and cross-validation, improving prediction reliability by 20% and reducing false positives CNN-based Image Classification with Tiny ImageNet Dataset, [GitHub] - Boston, MA

• Devised and implemented 5 different CNN architectures to classify images from the Tiny ImageNet dataset

• Utilized multiple image augmentation techniques and transfer learning to enhance classification accuracy

• Crafted Python script for data preprocessing, image labeling, organization, facilitating efficient model training

• Optimized CNN configurations, ranging from basic three-layer networks to intricate deep neural networks, achieving a significant boost in validation accuracy to 85% using a pre-trained ResNet50 model Automated Medical Report Summarization and Terminology Extraction, [GitHub] - Boston, MA

• Orchestrated preprocessing, optimized hyperparameters, and evaluated performance to achieve a 0.78 ROUGE score with BioBERT, although exploration of Pegasus - PubMed was undertaken

• Leveraged pre-trained model SciSpacy for NER, accurately extracting over 95% of medical terms from summaries

• Converted biomedical jargon into layman terms using Wikipedia for multi-word phrases, NLTK WordNet for single-word phrases with the ScispaCy model, enhancing accessibility and comprehension

Contact this candidate