Yugal Jain

About Yugal Jain

Experienced in building, Data Engineering (ETL) pipelines, end-to-end NLP pipelines and deep learning architectures- Proficient in Python, PyTorch, Sklearn, Pandas, Transformers, ElasticSearch, Flask, Docker, AWS, Spacy, NLTK- Have worked on supervised NLP problems such as Sequence Classification tasks(Aspect classification), Multi-Label Text Classification, Document Classification, Sentiment Analysis and Unsupervised Tasks such as Hierarchic Topic Modeling and Top2Vec.* Projects- Umeed : Advanced Language Analytics Tool(Delhi Police)• Experimented with NLP models to combat fake news, hate speech, and abusive content. Cleaned pre-existing data and trained Deep Learning model to detectabuse, traumatic and disturbing content in images and videos shared on socialmedia.• This is being undertaken to shed some light on the practical prospects of stopping the cycle of online crime, harassment and abuse.• Worked with multiple stakeholders to reduce the legal liabilities of models below the human level bias- Auto Content Moderation System• Developed an auto content moderation system to provide family friendly web shows and integrated with speech,text,and vision system which has ability to detectand mute abusive videos. Used video classifcation model to classify NSFW videos and CMU sphinx for Text to Speech(TTS) to detect abusive words in audioand replace it with beep.• Proposed a mode named SAFE MODE/FAMILY MODE. This mode uses our ACS(Auto Censoring System) to minimize explicit visuals and abusive languagefrom the content user wants to watch. Designed it in a way that it’s easily integrable with OTT platforms- Social Media Bias Free Bot• This bot can reply unbiasedly on comments posted by users on social media platforms like reddit,twitter or discord and will try to stop spreading racism,religious bias on these platforms through positive sentiment comments.• Trained GPT-2 based model on Jigsaw Toxic Comment Classifcation Dataset to detect toxicity in comments and AG-News Dataset to reply accordingto topic of comments posted on subreddits like politics, sports, technology etc and deployed on discord and reddit

Experience

1. Artificial Intelligence Engineer

   ValueSoft Info Services

   Aug 2025 — Present · Bengaluru, IN

   Client: Siemens Architected scalable RAG-based chatbot leveraging vector databases and semantic search to extract accurate answers from lengthy financial documents. Implemented enterprise-grade document processing pipelines with advanced chunking strategies, Azure Document Intelligence integration, re-ranking mechanisms, and entity extraction. Developed AIOps solutions for intelligent log parsing, system monitoring, and automated anomaly detection to improve operational visibility.

Education

• Guru Gobind Singh Indraprastha University

  Bachelor of Technology, Computer Science Engineering

  2017 — 2021