Natasha Tomar

Experienced ML engineer specialized in recommendation systems, deep learning, NLP, and AWS model deployment. Proficient in Flask for API development. Skilled in large language models, transformers, Elasticsearch, AWS, Docker, streaming data (Kinesis), PySpark model training, and various databases (DynamoDB, MongoDB, SQL, PostgreSQL). Proficient in ETL processes for data cleaning, normalization, preprocessing, and generating insights.

Highly skilled in developing advanced recommender systems across various domains, utilizing both content-based and collaborative filtering techniques. Proficient in information retrieval and recommendation algorithms, including cosine similarity, TF-IDF, and the universal sentence encoder. Accomplished in achieving high accuracy rates in recommending learning paths, jobs, and roles through the application of these algorithms.

Experienced in developing and training machine learning models for collaborative filtering, employing techniques such as ALS, NCF, and Two-tower architecture. Successfully created personalized recommendations based on user preferences and behavior, delivering impressive accuracy metrics. Demonstrated a precision of 0.90 and a recall of 0.85 in suggesting relevant learning paths to users, as well as an F1-score of 0.92 in recommending suitable job roles to candidates based on their skills and career aspirations.

Utilized reinforcement learning techniques, including multi-armed bandit and actor-critic algorithms, to optimize the learning-path recommendation process and enhance the system's ability to discover new items of interest.

Experienced in skill-role mapping, providing accurate recommendations for roles based on technical skills, industry, and preferences. Developed a model that combines universal sentence encoder embeddings and elastic search techniques to deliver tailored recommendations.

Proficient in developing semantic search systems using open-source search engines like Elasticsearch, enabling the design and implementation of efficient and effective search functionalities.

Extensive knowledge and practical experience with large language models (LLMs) such as GPT-3 and BERT. Conducted thorough experimentation and analysis to understand their capabilities and applications. Generated valuable insights on leveraging LLMs to improve natural language processing tasks.

Overall, a highly skilled and accomplished professional in developing recommender systems, with expertise in content-based and collaborative filtering, reinforcement learning, skill-role mapping, semantic search, and the utilization of large language models.

University: University Institute of Engineering and Technology CSJM University, Kanpur

Degree: Bachelor of Technology in Computer Science

Duration: August 2018 - May 2022

Cumulative Performance Index (CPI): 9.73/10