Devansh Singh, a Computer Science undergraduate at SRM University (CGPA: 8.91), has built a track record of high-impact ML solutions: 15% improvement in recommendation precision, 90% audio classification accuracy, and deployed systems processing hundreds of concurrent sessions. He is currently pursuing research on bias in Hindi/Hinglish toxicity classification.
Python, SQL, JavaScript, HTML, CSS
Git, Docker, MySQL, Streamlit, LangSmith, ChromaDB
TensorFlow, PyTorch, Keras, Scikit-learn, Hugging Face, LangChain, LangGraph, OpenAI API
NLP, Computer Vision, Recommender Systems, Time Series Forecasting
Statistical Modeling, A/B Testing, Feature Engineering, Production ML
Calibrating multilingual toxicity predictions with subgroup-aware threshold analysis and error slices for Hindi/Hinglish corpora.
Structured evaluation sets, guardrails, and prompt-level regression checks to reduce hallucinations before deployment.
Model drift alerts, response-latency bands, and ranking consistency metrics monitored like a daily newsroom dashboard.
B.Tech Computer Science Engineering
SRM University, Delhi-NCR · Aug 2023 – Jun 2027
CGPA: 0/10.0 — ML, Algorithms, OS, DBMS
Amazon ML Summer School 2025
Top 0 Nationwide
Neural Networks, GenAI, LLMs, RL, Causal Inference
Deep Learning (Infosys) · Big Data Bootcamp (MeitY) · Generative AI (Udemy)