Prakhar Agrawal

- Built RAG pipeline to generate visual art descriptions for 20K+ artworks using Multimodal LLMs - Developed and implemented evaluation methodology like classification of descriptions to whether they have subjective analysis - Accepted and presented findings at AI in Liberal Arts Symposium at Connecticut College - Gathered insights from 20+ vision-language papers to guide few-shot training

• Co-led the launch of an AI-powered recruiting platform adopted by 160+ candidates by engineering core infrastructure and features in React, Next.js, and Python; managed data using a PostgreSQL database • Boosted candidate-job match accuracy by 50% by implementing vector-similarity based resume ranking • Built & deployed sub-1100ms latency voice assessment agent using Deepgram, OpenAI & Cartesia on AWS Fargate • Ensured 100% database uptime by managing Supabase migrations and deploying secure authentication with Clerk

• Founding member of first-ever startup accelerator at college to assist startups in developing tech and raising capital • Created 10+ software engineering templates for mobile app, web dev, and web scraping, including a working stock trading app

• Authored documentation on High-Performance Computing Cluster for college website, aiding 200+ student and faculty researchers • Created parallelized scripts for GPUs to do ML tasks like regression & time-series using PyTorch, TensorFlow, Keras, & RapidsAI • Studied and Customized LLMs to make Math, CS, and other technical textbooks more accessible for students

Main TA for the COSC-211 Data Strucutres class.

Teaching Assistant for: - ECON 331 Advanced Macroeconomics - ECON 330 Macroeconomics

• Prepared material for board presentation, administered business operations with the CEO, and streamlined 5+ workflow systems • Collaborated with the engineering team to improve the product, and customize it for Police Departments across the US

• Customized LLM APIs to deliver personalized sports information to users, improving user engagement • Developed web scraping algorithms to gather real-time data from news websites, improving response accuracy • Implemented Web3 tokenization with blockchain to track and analyze user engagement

InMedia is a digital marketing start-up that aims to help local business to market themselves while providing paid social media marketing internships to high school and college students. The startup caters to more than 30 clients and provided virtual internships to 70+ students. Interviewed at “Kidsprenuer, business gyaan by young mentors”, an event showcasing start-ups of children below 18 years, organized by FICCI FLO, Indore chapter on Facebook Live, Jan’21. Published in National Daily Newapapers on 6th Jan’21 in “Free Press” (English) “Patrika Plus” (Hindi) and “Nai Duniya” (Hindi), and again covered on 10th Jan’21 & 12th June’21 in “Patrika” (Hindi).

Worked to optimise and reduce the costs of running waste management systems in Indore, the Cleanest City of India

Worked on a cyber security project to detect anomalies and fraud by mapping call records on Graph Databases on ArangoDB Oasis platform for the Crime branch, Indore.

During this Industrial Internship, Learnt Android App development, Web designing and data analysis. Successfully developed “Bank IFSC” app (live on google playstore), which enables user to find bank IFS Code of 4000+ bank branches of India.

Relevant Courses: Database Systems, Computer Security, Internet Networking, Electronics for IoT, Entrepreneurship

Honors/Awards: Outstanding Cambridge Learners' Award for World Topper in Math, Outstanding Cambridge Learners' Award for National Topper in 8 subjects, Gold Medal for Best All-Round Student, Trophy for Best in Co-Curricular Activities, Medal for Best Community Service project, Medal for Best Basketball Player, Prize for Best Hindi Debater

In this highly selective course comprising of only 17 students worldwide, students were taught how intelligent systems learn to do given tasks, and how optimization algorithms are used in Artificial Intelligence. Students were also introduced to Mathematical background of Machine Learning and they developed their own machine learning and deep learning programs.

The program taught by Professor Rajiv Gandhi, Professor at University of Pennsylvania, covered the mathematical foundations of Computer Science including topics in discreet mathematics such as systematic counting and methods of proof. The proof techniques included induction and basic graph theory to prove various theorems. Additionally, students were exposed to combinatorics and probability, set theory, advanced permutations and combinations, conditional probability, linearity of expectations, Big-O notation, cryptography, and planar graph coloring.

This course was designed to provide students an introduction to the field of finance, exploring the institutions, instruments and economics of financial markets. Starting with an introduction to economic models of the financial systems, students explored issues like how it creates value and how it fits into the larger macro economy. This course further progressed to discuss theories of asset pricing, exploring the efficient markets hypotheses and standard tools of valuation before turning to specific instruments as applications.

This program was designed for high school students provided global perspectives, holistic learning and ethical engagement through theme-based experimental learning in interdisciplinary areas – Law, Justice and Democracy; Business Management, Leadership and Entrepreneurship; Diplomacy, International Relations and Peace; Economy, Public Policy and Development; Liberal Arts, Culture and Humanities; Media Communication and Public Affairs; Architecture, Planning and Sustainability; and Banking, Finance and Accounting. The program included factory visit to Dawat Basmati Rice Factory.