Abhishek Malik

Engaged in an intensive machine learning program, applying advanced techniques to solve real-world problems and optimize diverse ML models.

• Applied supervised and unsupervised machine learning techniques using Python libraries including scikit-learn, NumPy, and pandas to address real-world problems.
• Built and optimized diverse ML models across multiple domains, consistently achieving accuracy scores above 95%.

Spearheaded the development of real-time face detection and recognition systems, integrating advanced ML algorithms for enhanced accuracy and data analysis.

• Implemented a real-time face detection system using Pandas, OpenCV, and NumPy, processing webcam input for data analysis and storage.
• Developed a facial recognition system leveraging the K-nearest neighbors (KNN) algorithm to accurately identify individuals.
• Identified and resolved critical issues in face recognition outputs, refining system performance to minimize errors and significantly enhance accuracy.

Modernized data infrastructure and improved scheduling efficiency through automated data analysis and professor collaboration.

• Modernized data infrastructure for minor subject analysis, transitioning from manual spreadsheets to automated SQL queries, reducing analysis time by 40%.
• Empowered faster schedule adjustments and enhanced data accessibility for academic stakeholders.
• Collaborated with professors to optimize semester scheduling efficiency and operational workflows.

Developed and refined Large Language Models (LLMs) by crafting and testing coding prompts, significantly boosting model accuracy and performance.

• Authored and curated over 200 coding prompts, rated 1-10, used to train LLMs for improved performance.
• Applied iterative retraining methodologies to boost LLM accuracy from 68% to 85% within a short timeframe.
• Assessed LLM performance by testing with custom-written prompts and classifying responses for accuracy against Python program outputs.

Led technical teams in Machine Learning and LLM development, fostering innovation within academic societies at Delhi Technological University.

• Led a team within the Inferno Society, driving the development and application of ML/LLM models.
• Co-headed technical initiatives for the Cultural Council DTU, promoting innovation in AI/ML technologies.

Organized major cultural and technical events, demonstrating leadership and project management skills within a university setting.

• Coordinated EngiFest '23, India's largest cultural fest, managing logistics and participant engagement for a large-scale event.
• Organized Smart India Hackathon '23, overseeing event execution and supporting participants through various stages.

Developed a comprehensive data analysis and visualization solution for credit card transactions and customer profiles using Power BI, SQL, and Excel.

Conducted in-depth sales data analysis using Pandas, NumPy, Matplotlib, and SQL to derive actionable insights for marketing strategy and business growth.

Developed an AI-driven fashion recommendation system utilizing ResNet50, NLP, and web scraping for enhanced customer engagement.

Developed a Retrieval Augmented Generation (RAG) based chatbot integrating Large Language Models (LLMs) for intelligent query handling using Python, LangChain, and Ollama.

Selected from over 100,000 applicants for Amazon's prestigious Machine Learning Summer School.

Maintained strong competitive programming ratings on Leetcode (1866) and Codeforces (1500+), showcasing advanced algorithmic and problem-solving abilities.

Successfully advanced through internal rounds of a national-level hackathon, demonstrating strong problem-solving and technical skills.

Achieved a top rank among 2 million candidates in India's highly competitive engineering entrance examination.

Achieved exceptional performance in Science and Mathematics Olympiads, ranking in the top 1% nationally at secondary level.

Qualified for the National Talent Search Examination Stage-1, recognizing academic excellence at a national level in Class 10th.