Navinkumar Palanisamy

I build

VIEW WORK R25
Navinkumar Palanisamy

Signal: Now

Building Finance Tracker Pro — a browser-based personal finance app for monthly cash flow, investments, expenses, portfolio holdings, goals, insurance, and a net-worth Overview dashboard.

Client-only on main: data in localStorage as one JSON document, with JSON export/import for backup and migration — plus light/dark theme and rich analytics views.

On GitHub

System Status: Online


class Engineer:
    def __init__(self):
        self.name = "Navinkumar"
        self.role = "Data Science Engineer"
        self.degree = "M.Sc. Embedded Systems"
        self.university = "Uppsala University, Sweden"
        self.edu_years = "2019–2021"
        self.focus = ["Deep Learning", "AWS", "GenAI"]
        self.location = "Stockholm, Sweden"

    def solve_problem(self, data):
        insights = self.analyze(data)
        return insights

// Professional Summary

Data Science Engineer with 4+ years of experience architecting high-performance ML solutions. Specializing in eye-tracking analytics and cloud-native deployments.

Currently leveraging LLM-driven agents and RAG workflows on AWS to revolutionize how attention data is interpreted. I bridge the gap between complex model research and scalable production systems.

// Education

Master's in Embedded Systems, Uppsala University, Sweden.

Execution Logs

Jan 2025 - Present

Cloud Intelligence R&D Team

Tobii AB | Stockholm

Prototyping GenAI & RAG pipelines using AWS Bedrock. Orchestrating multi-modal agents (Vision + LLM) to automate study insights.

GenAI AWS Bedrock RAG
July 2021 - Present

Data Science Engineer

Tobii AB | Stockholm

Optimized OpenVINO models (4x faster). Architected serverless AWS solutions (Lambda, DynamoDB) reducing costs by 12x.

OpenVINO AWS Lambda DynamoDB
Jan 2021 - June 2021

Master Thesis Student

Tobii AB | Stockholm

Model quantization & pruning for webcam eye-tracking. Reduced model size by 3.5x and latency by 36%.

Quantization Computer Vision

Tech Stack Protocols

AI & Deep Learning
PyTorch TensorFlow OpenCV ONNX OpenVINO CNNs Transformers
Generative AI & LLMs
LangChain RAG AWS Bedrock Agentic Workflows Vector DBs Prompt Eng Stable Diffusion
Cloud Infrastructure
AWS Lambda ECS SageMaker Docker CDK bash DynamoDB Bedrock
Core Languages
Python C++ SQL Bash Git

Deployed Modules

Cloud GenAI Prototypes

Automated attention data analysis using LLMs. Reduced manual reporting time by 80% via RAG workflows.

LLM Python AWS

Serverless Tracking

High-scale object tracking pipeline on AWS Lambda & DynamoDB. 12x cost reduction over EC2.

Serverless NoSQL

Edge AI Optimization

Real-time eye tracking on edge devices. Ported to OpenVINO with aggressive quantization.

OpenVINO C++

Beyond the terminal, the real world awaits!

View Off-Duty Logs