Soham Aggarwal

• Built a max-flow collateral optimization engine to minimize CCP cash margins across 1M+ clients. • Implemented Dinic’s algorithm for low-latency margin recomputation with sparse ownership graphs. • Reduced runtime by eliminating non-binding stock nodes which cut the active graph by 70%. • Generated per-client margin coverage, collateral utilization, and cash shortfall metrics risk analysis. • Currently being mentored in decentralized finance

Working closely with the leadership team at Interpro to: • Keep content fresh by updating page templates, copy, and media. • Improve visibility through effective SEO, including page content, meta tags, and metadata. • Review page design and embedded features to enhance user experience. • Monitor website performance (like load speed and core web vitals) and suggest improvements. • Track analytics to understand traffic, visitors, conversions, and the full customer journey.

• Worked with the research team to redesign the UI/UX of 200+ web pages • Increased website traffic by 53.4% through SEO and digital marketing. Raised speed by 36.8% • Managed online logistics and hosting for 20+ major events with 1,500+ attendees • Coordinated content updates & quality reviews across 10+ modules by completing 450+ revisions

• Led a 12+ programming team of LanchxLabs.ai on AI/ML development for patient-centric healthcare tools. • Utilized MERN Stack to build core features for Neuro86, a web app for neural-disease management • Improved app reliability and performance by reducing average page load by 40%, & API response time by 30% • Designed the patient medical-timeline system which helped reduce data-navigation time by 35% during internal testing.

• Developing a web-app for Neuro86 – an AI & ML powered tech startup project focused on managing the end-to-end journey of neural disease patients (a community for the same) – by utilizing Next.js & MongoDB. • Working under the supervision of Prof. Ramesh Raskar [MIT]. Involved from the planning stages and promoted to Programming Team Lead within 3 months. • Designed the logo and coded the patient medical timeline component.

• Engineered ML models (Scikit-learn) to process multi-omics data, and identified predictive biomarkers with 91% accuracy. • Developed algos to map tumor-immune interactions across 10k+ clinical records which reduced turnaround time by 35% • Paper published in the APJCC titled ”Dissecting the TME to decipher new immunotherapy targets using AI” • Currently working on the research paper

- Final grade for A1.1 - 95+/100 (under Ms. Virginia Nieto) - received a score corresponding to the highest level possible (95-100 marks) : SOBRESALIENTE 10 - Class rank 1

- Final grade for A1.1 - 96.75/100 (under Ms. Ramita Menon) - received a score corresponding to the highest level possible (95-100 marks) : SOBRESALIENTE 10 - Class rank 2

• Collaborated with the development team to integrate advanced language models and multi-regression algorithms for determining personality traits. • Assisted in designing and implementing prompt templates for user interactions. • Researched about the Big Five personality traits detection using Large Language Models (88% accuracy)

• Selected as one of 50 innovators from over 5,000 applicants to work on drone technology innovations, including drone docking stations. • Conducted research to analyze the competitive landscape and identify limitations in existing solutions.

Full Stack Software Developer & AI Engineer Resident with Full Scholarship! 14 advanced projects in 17 weeks.

• Built 5 AI-based projects, including TrackAI and QuestAI. • Utilized Next.js, React.js, NextAuth, MUI, TailwindCSS, Spline, CSS, JS, TypeScript, and integrated databases like Firebase & MongoDB.

Relevant Coursework: - Calculus -> MATH 221 & MATH 222 (Calc 1 & 2 - AP Credit), MATH 234 (Calc 3 - Tested Out) - Computer Science -> COMP SCI 200 (AP Credit), COMP SCI 300, COMP SCI 400 (enrolled) - Discrete Mathematics -> COMP SCI 240 - Linear Algebra -> MATH 340 (Tested out) - Probability Theory -> MATH 431 (enrolled) - Number Theory -> MATH 467 (enrolled) - General Education Classes (Gen Eds) -> AFROAMER 156, ANTRHO 105, COUN PSY 125, ENGL 153 (enrolled), COM ARTS 200 (enrolled)

- Selected as one of 50 students from all over India (Undergrad, Masters, and PhD students) - This school and workshop was held at IIT Delhi and focused on current techniques for achieving privacy and highlited some of the major open challenges in the field. - During the first two days of the program we covered foundational concepts of Cryptography, Differential Privacy, and Security. In the following two days, we had research talks outlining the real-world challenges in data privacy, societal and ethical issues, and how techniques from cryptography and differential privacy can be applied to address them.

• Completed CSCI P-14115 course (Introduction to Data Science with a Focus on Visualization) under Dr. Robert Roessler. • Gained proficiency in Python and Pandas for data visualization, web scraping, and filtering data using JavaScript. • Created websites using HTML to display visualizations of COVID-19 data. • Recognized as one of the top students in the official transcript.