Aditya Raj Pandey +91-7704943745
Software Engineer-AI/ML Intern #adityarajpandey770@[Link]
Bachelor of Technology ï [Link]/in/aditya-raj-pandey-815248230/
Indian Institute of Technology, Delhi § [Link]/Aditya23770
Education Specialization Institute Year
[Link] Engineering Physics IIT Delhi 2025
Scholastic Achievements
• Joint Entrance Examination (JEE) Advanced: Secured an AIR 4326 in JEE Advanced 2021.
• Joint Entrance Examination (JEE) Mains: Secured an AIR 6451 in JEE Mains 2021.
Work History
• Software Engineer - AI/ML | CheQ Digital Pvt. Ltd | Team CheQWisor | Bengaluru Sept.’25 – Present
– Data Annotation Project:- In-House Merchant NER Model Development & Deployment
∗ Led POC to build an in-house NER model, aiming to replace third-party merchant tagging vendors.
∗ Benchmarked SpaCy, GLiNER, & FinBERT by fine-tuning their softmax layers on a 10k dataset.
∗ Selected FinBERT as optimal; scaled its dataset from 10k to 30k and migrated tags from BIO to BILO.
∗ Achieved a 97% F1-score with the scaled FinBERT model and deployed it to the production assembly line.
– Current Projects: Output Guardrail Optimization & Session Sentiment Classifier
∗ Improving LLM output guardrails in CheQ Wisor to avoid false flagging using synthetic query evaluation.
∗ Building a prompt-based classifier to flag user dissatisfaction, miscommunication, or misinformation in chats.
• Data Analyst | Astra Business Services Pvt. Ltd. | Internship | Noida May’25 – Jun’25
– Automated workflows using Python, Pandas, and Selenium to convert Excel and scraped data into dashboards.
– Reduced manual workload by 50% by building a streamlined reporting pipeline for the VP, CEO, and DOO.
– Mapped BPO billing processes and automated dashboards to monitor key financial and operational KPIs.
Projects
• LLM-Powered SQL Chatbot: Generative AI, Prompt Engineering, Full-Stack Development | Self Project
– Built LLM-based chatbot using FastAPI to convert natural language into executable SQL database queries.
– Integrated Groq-hosted Llama 3 model with PostgreSQL, enabling AI-driven real-time query translation.
– Developed secure ReactJS frontend with formatted response display, logging, and error handling features.
• Fake News Detection Using Naive Bayes: Text Classification | NLP | Prof. Rahul Garg
– Implemented Bernoulli and Multinomial Naive Bayes with n-grams to detect fake news using LIAR dataset.
– Achieved 80% accuracy, improved F1-Score by leveraging contextual features like speaker attributes.
– Applied preprocessing techniques (stopword removal) to significantly improve data and model performance.
• Cyclone Preheater Anomaly Detection: Unsupervised ML | Anomaly Detection | DBSCAN | Self Project
– Analyzed high-frequency sensor data (370k+ entries, 3 years) using DBSCAN to detect time-series anomalies.
– Applied Min-Max Scaling, handled non-numeric values, tuned clustering via k-distance and silhouette scores.
– Detected anomalies in pressure and temperature signals; visualized system-wide failures and localized deviations.
• Movie Box Office Data Scraper: Web Scraping | ETL | Data Engineering | Self Project
– Developed a web scraper with BeautifulSoup to extract box office data for 3000+ movies from [Link].
– Implemented parallel processing via ThreadPoolExecutor for faster scraping of large datasets.
– Built an ETL pipeline to clean, transform, and load movie data into CSV files, integrating with TMDB metadata.
Technical Skills
Programming & Data Handling: Python, Java, JavaScript, MySQL, NumPy, Pandas, Matplotlib, Seaborn, Advanced
Excel, BeautifulSoup, ThreadPoolExecutor
Machine Learning & Agentic AI: Scikit-learn, TensorFlow, GPT-4, LangChain Agents, FAISS Vector Store, RAG,
Sentence Transformers, TF-IDF Vectorization, Prompt Chaining, LLM Evaluation, Naive Bayes, Logistic Regression,
KNN, Decision Trees, CNNs, Cosine Similarity, Sentiment Analysis, Feature Engineering, Hyperparameter Tuning
Tools & Platforms: LangChain, Streamlit, OpenAI API (GPT-4), Hugging Face Transformers, [Link] API
Positions of Responsibility
Logistics Coordinator (RendezvousX, Sept 2023 – Oct 2023):Led a dedicated 10-member team managing
comprehensive end-to-end logistics for 3 major cultural events, ensuring seamless and flawless execution.
Events Coordinator (PhySoc, Aug 2023 – Apr 2024):Organized events including Intern Insights, faculty-student
meetups,[Link] cross-functional marketing & publicity team to enhance engagement & collaboration.
Events Executive (Tryst’23, Dec 2022 – Mar 2023):Led initiatives securing sponsorships from Quickheal and Neovel,
orchestrated 6 successful events with CXO-level collaboration, and achieved 700+ registrations