Saubhagya Singh
Jr. Data Scientist / AI Engineer @ Q3 Technologies | B.Tech IT @ NIT Kurukshetra | AWS Bedrock, RAG, LangGraph | Gen AI
Chandigarh
2+ years of work experience
Feb 20251 yr 9 mosPresent
Gurugram, Haryana, India
NeuroHub:
- Built a hybrid LSTM + 2D CNN model capturing short- and long-term ECG patterns, improving seizure detection accuracy by 84%
- Built a pipeline to localize and flag seizure events and signal artifacts for UI visualization
TransElite Recommendation System:
- Built and Dockerized deployment of a 52,000+ line recommendation engine combining LightGcn, collaborative filtering and content-based models via RRF and EMA-weighted ensemble
- Engineered a feedback loop using dispatcher accept/reject signals to update ensemble weights, reaching a 91.4% acceptance rate over 3 months
- Built the DB connectivity layer for dispatcher feedback capture into the recommendation loop
Auraclinical AI:
- Built an end-to-end LLM pipeline (Claude Sonnet 4.6 + Haiku 4.5 on AWS Bedrock) converting sponsor documents into ICH E6(R2)-compliant clinical trial protocols using hybrid retrieval (ChromaDB + BM25) and agentic verification
- Architected a pre-flight Intake Agent and a two-stage QC system (structural + LLM hallucination checks) with keep/flag/regenerate/escalate decisions
- Set up AWS Bedrock batch inference infra (IAM, S3), cutting inference costs ~60% and improving pipeline speed ~30%
Keypath Educational Platform:
- Designed an OCR + pdfplumber extraction agent (Azure Document Intelligence).
Jan 20244 mosApr 2024
Mathura
- Spearheaded the development and maintenance of scalable web applications using React and Redux, ensuring robust interfaces
- Reduced code redundancy by 46%, enhancing maintainability and increasing readability by 30% using Redux
- Boosted operational efficiency by 33% through efficient order processing and decreased error rate
Dec 20236 mosMay 2024
Hyderabad
- Upgraded UI usability by 30%, focusing on profile and address sections for enhanced visual appeal and responsiveness
- Strategically integrated backend APIs, ensuring 99% device compatibility and responsiveness
- Engineered and delivered 7+ new features on time, resolving 10+ critical technical issues, enhancing user satisfaction by 25% and accelerating product release cycles by 15%
Dec 20236 mosMay 2024
– Delivered an admin-management frontend and strengthened backend API integration across device types, improving UI
usability by 30%.
Saubhagya’s AI Stack
Education
Bachelor of Technology, Information Technology
National Institute of Technology, Kurukshetra, Haryana
20214 yrs2025
Class 12th, CBSE Board
Asian Public School Gurgaon
20191 yr2020
Join Saubhagya and 10k+ other professionals on
Tal is the app where bosses meet engineers & builders directly, without middlemen
Download Tal Boss