Headline

Aspiring Data Engineer & Data Scientist | Python · ML · NLP · SQL · LLM Agents | IEEE Published | 3 Internships | B.Tech CSE @ Amity | Open to Opportunities

About

Final-year B.Tech CSE student at Amity University, Greater Noida (CGPA: 8.62) with hands-on experience in Data Engineering, Data Science, and AI/ML systems. Over the past year, I've completed 3 internships — Data Engineering at Cognifyz Technologies, Data Science at Oasis Infobyte, and Java Development at CodSoft — where I built ML pipelines, designed ETL data workflows, and performed EDA on datasets with 10,000+ records. My projects span the full DS/DE spectrum: Built a CNN-based Deepfake Detection System using TensorFlow trained on the FaceForensics++ dataset, achieving 90%+ accuracy with real-time inference via a Streamlit app Developed an NLP spam classifier using Naive Bayes & SVM on 5,000+ samples with 88–90% accuracy Engineered DataGuardian for the Meta × Scaler OpenEnv Hackathon — a reinforcement learning environment where an AI agent audits and fixes messy datasets, scoring 0.946/1.0 on the hardest task and outperforming the GPT-4o baseline of 0.82 — deployed live on Hugging Face Spaces with FastAPI + Docker I also have a published research paper on IEEE Xplore on the topic: " DeepfakeBench: A Comprehensive Benchmark of Deepfake Detection" — a milestone I'm proud of as an undergraduate. Tech stack: Python · SQL · TensorFlow · Scikit-learn · FastAPI · Docker · OpenAI API · Streamlit · Pandas · NumPy · Power BI · Tableau · Hugging Face Actively seeking entry-level roles or internships in Data Engineering, Data Science, or AI/ML. Always open to collaborating on interesting projects too. Feel free to connect or reach out — let's build something meaningful with data!

Location

Greater Delhi Area

AI stack

hugging face

langchain

ollama

openai

Work experience

Participant — META OpenEnv Hackathon at Meta, Mar 2026 - Apr 2026 · 2 mos, Bengaluru, Karnataka, India, DataGuardian — Real-World Data Quality OpenEnv DataGuardian is a reinforcement learning environment built for the Meta × Scaler OpenEnv Hackathon (April 2026), where an AI agent learns to audit and fix messy datasets like a senior data engineer. The Problem: Data engineers spend 60% of their time cleaning dirty data. Nulls, type errors, duplicates, outliers, and format errors cost companies millions in lost productivity. No reinforcement learning environment existed for this domain — until now. How It Works: The agent inspects one data cell at a time across two realistic datasets — customer orders and employee records (30 rows total, 20+ unique data quality issues). For each cell it chooses from 11 possible actions spanning detection and fixing. Correct decisions earn reward. Wrong decisions earn zero. The agent learns purely from this feedback loop across thousands of episodes. Three Tasks: Task 1 (Easy) requires detecting obvious issues in 4 rows. Task 2 (Medium) requires detecting AND applying the correct fix across 8 rows. Task 3 (Hard) is a full pipeline audit across both datasets shuffled together — even GPT-4o scores only 0.82. The built-in auto-demo agent scored 0.946. Technical Highlights: OpenEnv-compliant with Pydantic typed models (step/reset/state API) FastAPI backend with 9 endpoints including /metrics and /leaderboard LLM agent with chain-of-thought reasoning via OpenAI client Dense partial-credit reward function — not just binary win/lose Dockerized and deployed live on Hugging Face Spaces Interactive demo UI with AI Auto-Demo button All automated hackathon checks passed ✅

Data Science Intern at Oasis Infobyte, May 2025 - Jun 2025 · 2 mos, India

Data Engineering Intern at Cognifyz Technologies, Apr 2025 - May 2025 · 2 mos, India

Student- Class Representative at Amity University, Greater Noida Campus, Feb 2024 - Jul 2024 · 6 mos

Java Programming Intern at CodSoft, May 2024 - Jun 2024 · 2 mos, India

Education

B.Tech(CSE), Computer Science, Amity University, Greater Noida Campus

Completed 10+2, PCM, Somerville School - India