Shaurya Suman.

Engineering

Projects, stack, and CV — for recruiters and academic reviewers, not clients.

Research internship — IIT (BHU), Varanasi

Summer Intern, Dept. of Computer Science & Engineering · under Dr. Bidyut Kumar Patra, Associate Professor · July–Sept 2025

  • Implemented recommendation algorithms — user-user and item-item collaborative filtering, and matrix factorization — in Python on the MovieLens 100K dataset, with an 80/20 train-test evaluation setup
  • Built reusable data-processing pipelines so experiments could be configured and repeated without rewriting code
  • Handled the cold-start case — users with no rating history — with a popularity-based fallback, and generated diverse Top-N recommendations with genre-level filtering
  • Documented the codebase and results, and presented technical decisions and trade-offs to the faculty research team
PythonCollaborative FilteringMatrix FactorizationMovieLens 100K

AI Research & Intelligence Engine

Personal project — built independently, not for an employer or client

A hundred-process research pipeline that runs long, unattended research jobs. FastAPI backend, with ChromaDB and sentence-transformers handling vector search over research sources, and a local model served through Ollama (qwen2.5:7b) doing inference — nothing gets sent to a third-party API.

The part worth explaining is the crash recovery. Every process checkpoints its state to SQLite as it runs, so a crash — or a deliberate stop — doesn't lose progress: on restart, the pipeline reads the last checkpoint and resumes from there instead of starting the whole run over. Data passed between stages is validated against Pydantic schemas, so a malformed intermediate result fails immediately and visibly instead of silently corrupting the next stage.

FastAPIChromaDBsentence-transformersOllamaSQLitePydantic

Student Feedback Analysis

Final-year project, MAKAUT · 2026

A full-stack tool that lets faculty upload student feedback as a CSV and instantly see it classified by sentiment — classical NLP, not a language model: TF-IDF vectorization feeding a logistic regression classifier.

  • FastAPI REST backend for upload, processing, and results retrieval, with validation so a malformed CSV fails with a clear error instead of crashing
  • React dashboard for upload and interactive visualization of the sentiment distribution
  • Documented for handoff, so the department can keep running it without my involvement
PythonFastAPIReactScikit-learnNLTK

netops-toolkit

Ongoing, early-stage

A repo for learning network automation from the ground up — Linux administration, permissions, and Python fundamentals so far, with Netmiko-based device automation planned next. Public and unpolished on purpose: it's where the fundamentals get built before the automation layer goes on top.

PythonLinuxNetmiko (planned)

Education

Degree
B.Tech, Information Technology
Institution
MAKAUT, West Bengal
Period
2022 – 2026
CGPA
8.16 / 10
Coursework
DSA, DBMS, Operating Systems, Computer Networks, Software Engineering

Skills

Languages

PythonC/C++JavaScriptSQL

Backend & APIs

FastAPIDjango/DRFNode.jsREST API design

Frontend

React

Data

PostgreSQLSQLiteMySQLRedisChromaDB

Infra & tools

DockerCeleryGitLinux/UbuntuVS Code

ML & NLP

scikit-learnPandasNumPyNLTKsentence-transformersOllamaLangChain (learning)

Networking

Netmiko (in progress)

Fundamentals

DSA (250+ problems, C++)OOPDBMSOperating SystemsComputer Networks

Recruiting, or want the full picture?

WhatsApp or email both work — the CV above has the same detail in one document.

Chat on WhatsApp