Hi, I'm Smyan
I'm a 3rd year Computer Science student at Northeastern University, passionate about building software and solving problems.
Experience
Machine Learning Engineering Co-op
- • Engineered a spaCy-based Natural Language Processing (NLP) chat assistant with a custom Named Entity Recognition (NER) pipeline from scratch, providing automated insights on experimental data, process documentation, and model optimization strategies for workflows across 5 distinct manufacturing processes
- • Developed, automated, and deployed a data cleaning and predictive modeling pipeline using XGBoost, Random Forest, and SVR on 9+ time-series lab datasets (1,000+ datapoints each) for an $11 billion polysaccharide data analysis project
- • Implemented SHAP explainability analysis and EconML causal inference to identify highest-impact factors across datasets, ensuring model interpretability and consistency
- • Researched and implemented MLflow experiment tracking and Data Version Control (DVC) for reproducible ML workflows; presented proof-of-concept to modeling teams across 4 sites
Teaching Assistant – Foundations of Data Science
- • Facilitated 6+ office hours per week, directed project meetings, and proctored labs/exams to assist students with assignments and aid them in understanding data analysis, linear algebra, statistics, and machine learning concepts
- • Graded 90+ assignments per week, providing useful feedback
Director of Operations
- • Secured Anthropic as a sponsor for two large-scale 24-hour hackathons, negotiated Claude API credit allocations for winners
- • Coordinated demos and workshops with Claude Student Ambassadors and student developers to drive adoption and showcase capabilities across the Northeastern developer community
Founding Member and Vice Chair
- • Directed the planning and execution of two large-scale 24-hour hackathons with 40+ attendees, coordinating budgeting, sponsorships, judging, workshops, and cross-club collaborations across 6+ student organizations and 3 local startups
- • Facilitated communications between student software engineers and local startups to develop impactful real-world products
Co-Founder and Vice President
- • Spearheaded an initiative to establish an interdisciplinary undergraduate research group focused on the intersection of computer science, data science, and the life sciences, facilitating communication between members and advisors
Tools I Use
Projects
NewsFactChecker
Designed a probabilistic NLP fact-checking model on 10K+ statements from the LIAR dataset using a 6-class Bayesian softmax classifier, applying Hamiltonian Monte Carlo with JAX to sample 5,000 posterior weight estimates. Implemented semantic feature extraction using SentenceTransformer embeddings; benchmarked posterior-averaged credibility scores against logistic regression baselines using AUC and threshold-optimized F1.
Guardrails: Atomic
1st Place - 2025 AI Agent HackathonArchitected an AI-powered full-stack formal verification platform in Next.js/TypeScript that iteratively generates mathematically verified code from natural language inputs, winning 1st place at an AI Agent Hackathon. Engineered a YAML-to-Z3 conversion pipeline with automated counterexample generation for program verification.
OpenLegislation
1st Place - HackHarvard 2024 Open Data TrackBuilt a web application that simplifies congressional legislation for general audiences, enabling users to search, read, and understand active bills without legal expertise, winning 1st place at the HackHarvard 2024 on the Open Data Track. Implemented semantic search using OpenAI embeddings to surface relevant bills from natural language queries, and integrated GPT-based summarization to translate legal language into plain English.
HealthSync
Developed a mobile application that analyzes user health data and health-related journal entries to provide users with AI-powered health analysis and recommendations.
Second Sight
Developed an AI-powered journaling application where users can log their mood for the day, create journal entries, and view entries over time. Leveraged Gemini AI to analyze entries and mood trends to assist users in understanding themselves better.
Hoo Wants A Degree
Developed a degree planner web application to aid University of Virginia School of Engineering students in creating AI-generated four-year degree plans.
Stocks Simulator
Developed a full-stack Java application using MVC architecture for users to create, update, and maintain stock portfolios. Incorporated algorithms to calculate stock gain/loss, moving averages, and portfolio rebalancing with real-time data for 1000+ stocks.
Contact
Feel free to reach out if you'd like to connect or collaborate!