Soham Sinha
AI Systems Engineer • CS & Math @ UMass Amherst
Designing and deploying intelligent systems at the intersection of machine learning and software engineering, from multi-step LLM workflows and fine-tuned models to scalable APIs and production-grade infrastructure.
Featured Projects
Building intelligent systems and scalable infrastructure. From ML research to production deployments.
What I Build
AI/ML systems, full-stack AI products, and impact through teaching and collaboration.
- •Multi-step AI agentic workflows and LLM-powered applications
- •End-to-end ML pipelines: data cleaning, training, and evaluation
- •Fine-tuning and optimizing models (e.g. BERT) for real-world accuracy
- •FastAPI and Docker for scalable, production-ready services
- •Chrome extensions and web apps with React and Next.js
- •Conversational interfaces and data visualization powered by LLMs
- •Real-time pipelines and collaborative features
- •Cross-platform automation (e.g. desktop workflows, Apple Shortcuts)
- •Teaching and mentoring: UCA for 300+ students in introductory CS
- •Team leadership: led intern teams on PDF parsing and data extraction
- •Clear communication and collaboration with instructors and stakeholders
- •Analytical, research-oriented approach with a strong mathematical background
Experience
Building scalable systems and shipping AI products at top companies and startups.
Machine Learning Intern
Finarkein Analytics Pvt. Ltd. • Remote
Jul - Aug 2025
- •Enhanced text classification model accuracy from 67% to 90% by cleaning 100k+ data points, experimenting with tokenization, and fine-tuning BERT variants with Hugging Face
- •Engineered ML models from scratch, optimizing the full pipeline from data ingestion to evaluation—reduced model training time by 50%
- •Deployed an email extraction API with FastAPI and OpenAI APIs, improving operational efficiency
Software Engineering Intern
A79.ai • Remote
May - Aug 2025
- •Developed and deployed AI workflows for pharma leaders (Pfizer, Merck): deep research extraction agent for clinical trial research and forecasting agent for drug NPS
- •Delivered scalable, client-facing AI solutions that streamlined research and forecasting, simplified complex calculations, and enhanced data-driven decision-making; used Docker for deployment
AI Intern
Brijbasi Fire Safety Systems Pvt. Ltd. • Mumbai, India
Jun 2023
- •Collaborated with industry professionals to assess applications of AI and physics in firefighting
- •Delivered a report guiding the company on opportunities and challenges of integrating AI/ML into the industry
Intern
Rubix Data Sciences Pvt. Ltd. • Remote
Jul 2023
- •Led a 3-member intern team to build a PDF parsing program using Python (OpenCV and PyTesseract) and Java
- •Achieved 95% accuracy and reduced manual data entry, with multi-format information extraction
Undergraduate Course Assistant
Manning College of Information and Computer Sciences, UMass Amherst • Amherst, MA
Sep 2025 - Present
Supporting CICS 110 — introductory computer science
- •Supporting 300+ students in foundational CS concepts, programming fundamentals, and problem-solving
- •Host office hours for debugging, project ideas, and concept support
- •Collaborate with instructors and other UCAs to refine course materials and create engaging resources
About
Analytical and research-oriented—passionate about AI, natural language, and practical solutions to meaningful problems.
I'm a Computer Science and Mathematics double major (with a Business minor) at UMass Amherst, building a strong foundation in DSA, system design, computational proofs, and linear algebra. I'm passionate about AI, natural language, and creating practical solutions that deliver real-world impact.
I've built and shipped AI workflows at startups—from improving text classification at Finarkein and deploying agentic solutions for pharma at A79.ai, to leading a team on PDF parsing at Rubix. I also support 300+ students as an Undergraduate Course Assistant for introductory CS, and I love teaching and mentoring.
I'm seeking a summer internship in technology, specifically data and computer science. When I'm not coding, you'll find me on the court or field—I play basketball, cricket, table tennis, badminton, chess, and pickleball—or contributing to open source and side projects.
What I Value
Growth Mindset
Always learning. Curious about new tech, papers, and better ways to build and ship.
Integrity
Doing the right thing. Clear communication, respect for others, and honest collaboration.
Commitment
Seeing projects through. Building things that matter and delivering measurable impact.
Skills & Technologies
Languages
AI/ML
Backend & APIs
Frontend
Other
Let's Build Together
Looking for collaborators on ambitious projects. Reach out if you're working on something interesting.
Prefer email directly?