CV
Professional Summary
Experienced Machine Learning Engineer with 5+ years of hands-on experience in AI and NLP. Led award-winning personalized recommendation systems at TIME magazine, achieving a 42.97% open rate. Skilled in collaborating on innovative product strategies with executives and partners. Seeking to advance AI research through a Masters/Ph.D. program.
Research Intrest
- Trustworthy AI (Interpretability, Fairness, Robustness, and Privacy); LLM, RAI (Responsible AI), Local Explainable AI, NLP in Media and Medicine
Education
- Bachelors of Science in Computer Science and Information Technology (BSc. CSIT), Tribhuvan University 2016-2020
Work Experience
- Sr. Machine Learning Engineer at Fusemachines, Nepal Mar 2020 - Present
- Sr. Python, Data Science, AI/ML Trainner at Broadway Infosys Aug 2022 - Present
- AI Engineer at OYA INC Apr 2019 - Sept 2019
Skills
Languages: Python (5 years), C (6 months), C++ (6 months)
Libraries: Numpy Pandas, Matplotlib, Plotly, Seaborn, SciPy, Scikit-learn, NLTK, Tensorflow, OpenCV, FastAPI
Cloud Platforms and Services MLOps: Google Cloud Platform(GCP): VertexAI, Compute Engine, Cloud Run, Cloud Function, Cloud Build, Triggers, Load Balancing, Cloud Schedulers, GCS Bucket, Cloud Monitoring, Discovery Engine, Gemini, Datastores, Vector Search/Matching Engine, Agent Builder Tool.
Database: SQL, BigQuery (BQ), MongoDB
Version Control: Git, GitHub
ML Algoirthms: Linear/Logistic Regression, Decision Trees, Random Forest, NB, SVM, KNN, KMeans, PCA
Deep Learning Architectures: MLP, DNN, CNN, RNN, LSTM, AutoEncoders, Transformers
Soft Skills: Leadership, Teamwork, Communication, Presentation skills, Problem-solving
Others: A/B testing
Academic/Personal Projects
Nepali Cash Detection and Recognition (2019 - 2023)
Detecting and recognizing Nepali cash using InceptionV3 (open sourced)
- Developed and compared different Deep Learning Model to classify and recognize Nepali Cash/Bank Notes using TensorFlow and CNN + Transfer Learnning with 94% accuracy on the test and validation set
Disease NER on Clinical Data (2022)
NER to detect disease form clinical text
- Configured transformer model to detect disease names from Bio-bert clinical text and achieved 95% accuracy
JobsForSkills (Final Year Project, 2020)
Scraps recommended jobs for users based on their skills on the platform
- Prepared over 2400 job descriptions for 15 tech roles using web scraping (open-sourced dataset in Kaggle) and developed a classification-based job recommendation system with 84% accuracy