Vigneshwararaj

Computer Science Engineer

Vigneshwararaj Sathianarayanan. M.Sc. student in AI and machine learning at Linköping University.

Engineer first

Two years building backend systems at NielsenIQ. One engine I built cut ticket resolution from 2 days to real time.

I read hard signals

Machine sensors. WiFi waves. Ancient Tamil carved in stone.

1st prize, Global Hackfest

Innovation Award at NielsenIQ's Global Hackfest, 2024.

Looking for a thesis

A master's thesis in AI or ML, from January 2027. Let's talk.

Work

What I have built, on my own and in teams, from industrial sensor data to AI that runs on a laptop.

My local AI assistant

Personal project, ongoing, 2026

A local AI assistant on a 4 GB laptop GPU. It talks to tools through MCP servers I wrote, so it can read files, open apps and search the web without sending my data to the cloud.

View project
The VR Office panel showing a grammar fix, with changed words highlighted and Apply and Undo buttons

AI inside LibreOffice

Personal project, 2026

A LibreOffice extension that rewrites text, writes Calc formulas and answers questions about your document, using a small model running on your own laptop.

View project
Illustration of a floodlit cricket stadium at night, with glowing lines tracing shots across the field

What wins an IPL match

Sports analytics project, Linköping University, 2026

A real-time insights system built on ball-by-ball IPL data. It tracks win probability, finds which phases decide a match, and recommends who should bat or bowl next.

View project
Close-up of a steel deep groove ball bearing with its balls visible

A data and ML platform at SKF

Summer job, machine learning and data engineering, SKF Vertevo, 2026

I built a unified ingestion framework on Microsoft Fabric, a medallion architecture, an AI agent that writes its own file parsers, and a model that finds test failures.

View project
A synthetic training image from this project, with Vattezhuthu text on a stone texture

Reading Tamil script from stone

NLP course project, Linköping University, 2026

I fine-tuned Qwen2.5-VL to read Tamili and Vattezhuthu, two extinct Tamil scripts, from images of stone inscriptions. It scored 95.4 BLEU on clean images and 63.0 on images made to look like worn stone.

View project
The robot in my simulated office in RViz, with the costmap's personal-space field around a person

A robot that gives people room

AI robotics course project, Linköping University, 2026

A simulated TurtleBot4 that keeps a safer distance from people. I wrote a custom Nav2 costmap layer in C++ and trained a Soft Actor-Critic agent to adapt its personal-space settings as it moves.

View project
Sentinel-2 satellite image of a city and the land around it

Teaching CLIP to read satellite images

Deep learning group project, Linköping University, 2026

Fine-tuning OpenCLIP with LoRA to classify satellite images into categories it never trained on. I designed and built the team's second semi-supervised approach, which scored 69.89% on the hidden test set.

View project
Comic-style illustration of a phone on a table, with sound rising from it and splitting into separate coloured ribbons

Pick your conversation

Group project course, Linköping University, in progress, 2026

A phone on a table that separates several conversations happening at once, to help people with hearing loss follow the one they choose. I own speaker diarization and separation.

View project
Comic-style illustration of a magnifying glass pulling a few glowing speech bubbles out of a dark crowd of them

Finding software needs in Reddit

Text mining project, Linköping University, 2026

A three-stage NLP pipeline that sifts about 4,300 Reddit posts and comments for real software needs, separating solvable workflow problems from general venting.

View project
The Cinemate landing page with the headline "Track films you've watched. Tell your friends what's good."

A social network for film lovers

Group project with Shaurya Thakur, Linköping University, 2026

A full-stack web app where people rate and review films and shows, keep watchlists, follow friends and talk in threads. Built as a two-person course project and deployed online.

View project
Comic-style illustration of WiFi waves spreading from a circuit board through a room of people in silhouette

Counting people with WiFi signals

Bachelor thesis, Amrita Vishwa Vidyapeetham, 2023

A privacy-friendly way to count people in a room using WiFi Channel State Information instead of cameras. Phase-rotation augmentation took an LSTM from 49% to 67% accuracy.

View project
Comic-style illustration of supermarket shelves whose products turn into branching lines of data

Two years building at NielsenIQ

Software Engineer, Platform API, NielsenIQ, 2023–2025

I developed features for the Discover application, and worked on proofs of concept and R&D, including integrating LLMs into Discover. One engine I built cut ticket resolution from 2 days to real time.

View project

Also built

Journal

Notes from the build: what I tried, what broke, and what I learned.

All entries

Web search without an API key

Local assistant

Teaching a 3B model to call tools

Local assistant

A chat model in 4 GB of VRAM

Local assistant
Illustration of Vigneshwararaj on a rock by the sea, with the sage Thiruvalluvar behind him and the Thiruvalluvar statue at Kanyakumari in the distance

About

I'm Vigneshwararaj Sathianarayanan, a Computer Science Engineer. I spent two years as a software engineer at NielsenIQ in Chennai, then moved to Sweden for a master's in AI and machine learning.

This summer at SKF in Gothenburg, I built a data platform on Microsoft Fabric, an AI agent that writes its own file parsers, and a model that finds test failures. I like problems where the data is messy and the right approach is still open.

Experience

  • SKF Vertevo, machine learning and data engineering2026
  • NielsenIQ, software engineer2023–25

Education

  • Linköping University, M.Sc. Computer Science, AI and ML2025–27
  • Amrita Vishwa Vidyapeetham, B.Tech. Computer Science2019–23

Awards

  • 1st prize, Innovation Award, NielsenIQ Global Hackfest2024
  • Rising Star Impact Award, NielsenIQ2024
  • 2nd prize, Adventure Day Hackathon, Azure Monitoring2024
  • Finalist, Smart India Hackathon2022

Skills

Languages: Python, SQL, MATLAB, Java, C++

Machine learning: PyTorch, TensorFlow, time-series, anomaly detection, LLM fine-tuning, RAG and agents, reinforcement learning

Data and platforms: Apache Spark, Delta Lake, Microsoft Fabric, Azure, Snowflake, Git, Linux, ROS 2

Let's talk

I'm looking for a master's thesis in AI or ML, from January 2027, anywhere in Sweden.

vignesh181201@gmail.com