Headline

Founding Team @Weaver | Data Sciences | Scalable Agentic Intelligence

About

I’m a Data Scientist focused on building production AI systems that combine machine learning, LLM agents, and large-scale analytics. At Weaver Finance, I work on AI-native data platforms and intelligent analytics systems. I built Watson, a BI agent that enables natural-language querying across 1300+ tables, and designed a feature engineering pipeline that consolidated fragmented data sources into production-ready ML features. Previously at Probo, I developed ML and analytics systems used by millions of users. My work included LLM-powered trend detection, time-series forecasting models, fraud detection pipelines, and analytics systems that uncovered ₹2.5Cr+ in financial impact. I enjoy working on problems at the intersection of: • Machine Learning systems • LLM agents and RAG architectures • Data platforms and analytics infrastructure • Time-series modeling and predictive systems I’m particularly interested in roles where I can build intelligent systems that move beyond experimentation and deliver real-world impact.

Location

Noida

AI stack

claude

langchain

langsmith

manus

notebooklm

opencode

Work experience

Data Scientist at Weaver, Sep 2025 - Present · 1 yr, Mumbai, Maharashtra, India, - Part of the founding tech team in an AI first lending platform - Revolutionizing credit lifecycle, from initiation to evaluation to collection - Building one of a kind SDD first tech team

Data Scientist at Probo, Nov 2024 - Sep 2025 · 11 mos, Gurugram, Haryana, India, - Built an agentic chatbot with 12+ data ingestors using LangGraph, outperforming SOTA tools (e.g., Perplexity) for statistical query resolution. - Designed AskProbo, an LLM-powered trend detection system for tweets/events, generating 400K+ impressions and auto-answering 3.5K+ queries on X. - Developed Drishti, an event insights engine with caching-aware modular APIs, scaled seamlessly to millions of users via distributed message queues. - Created a time-series trader forecasting model (MAPE: 11%) that consistently delivered bounded daily predictions. - Engineered an analytics pipeline uncovering ₹1.5Cr+ dormant user funds, enabling targeted resurrection of high-value users. - Built fraud detection workflows to identify unfair event advantages, recovering ~₹1Cr in misused funds. - Automated an event creation pipeline ingesting 4K+ articles/day, with 90%+ accuracy in predicting future trending topics.

Member of Technical Staff (Data Science) Intern at GeeksforGeeks, Nov 2024 - Nov 2024 · 1 mo, Noida, Uttar Pradesh, India, - Worked on Data Science and Machine Learning related articles. - Wrote close to a dozen high impact new articles for the website. - Reformed and enhanced various existing articles on the topics of Machine learning, Deep learning, NLP, etc.

Summer Intern at Adobe, May 2024 - Jul 2024 · 3 mos, Noida, Uttar Pradesh, India, • Revolutionized service desk quality audits and triaging systems, slashing manual efforts by 10x. • Engineered and deployed advanced models, achieving 98-99% accuracy on real-time data. • Leveraged NLP, transformers, and large language models for impactful global automation.

Data Scientist Intern at Cornerstone Solutions, Jan 2024 - Apr 2024 · 4 mos, Indore, Madhya Pradesh, India, • Constructed a unique dataset with 800+ images for industrial structure analysis through web scraping. • Implemented ML and image processing models, attaining over 93% accuracy and a 64% map50 score. • Innovated camera feed automation methods, ensuring minimal delay, increasing real time accuracy by 30%. • Co-authored a research paper on vegetation detection and received a letter of recommendation.

Education

Bachelors of Technology, Computer Science, amity University

Class 12th CBSE, Science with Mathematics and Computer Science, Vishwa Bharati Public School Noida

Public links

github: https://github.com/om-shreyas

leetcode: https://leetcode.com/user1721IM