INITIALIZING...
0%
Available for work · AI/ML Engineer

Sree Harshan

Transformers. Kubernetes. Real-time systems.
Building intelligent infrastructure that ships.

YOLO
K8s
Transformer
FastAPI
VLM
0 Production Systems
0 Tech Domains
0 Core Skills
scroll

Who I Am

I'm an Associate Software Engineer at AvenirVista Tech Solutions, where I build production-grade AI infrastructure for global automotive clients. My work lives at the intersection of deep learning and systems engineering — transformer models deployed in real time, pods that spin up on demand, and dashboards that pulse with live hardware data.

B.Tech in AI & Machine Learning from Kongu Engineering College, Tamil Nadu. I care about systems that work under pressure.

LocationErode, Tamil Nadu
EducationB.Tech AI/ML · 7.74 GPA
InstitutionKongu Engineering College
RoleAssociate Software Engineer
CompanyAvenirVista Tech Solutions
Status● Active

Tech Stack

⌨️

Languages

Python JavaScript Java Bash
🧩

Frameworks

Node.js FastAPI Express.js HuggingFace Keras PySide6
🧠

AI / ML

LLMs VLMs YOLO Object Tracking Transformers
☁️

DevOps

Docker Kubernetes Nginx Jenkins CI/CD
🔧

Tools

Git Postman Linux Raspberry Pi

Work Experience

Associate Software Engineer

AvenirVista Tech Solutions · Coimbatore
Aug 2025 – Present
  • Trained an encoder-decoder transformer for real-time vehicle scenario prediction (TTC & TTI), deployed for a major global automotive client.
  • Built a backend that dynamically spins up Kubernetes pods on client connect and tears them down on disconnect, efficiently serving concurrent users.
  • Built a real-time CI/CD dashboard with live Raspberry Pi data feed, FastAPI backend, PySide6 desktop UI, and Jenkins integration.
  • Implemented a Kalman filter-based multi-object tracker for real-time vehicle tracking across video frames.
TransformersKubernetesFastAPIKalman FilterJenkins

Selected Work

01

Transformer-based Vehicle Scenario Prediction

Encoder-decoder transformer classifying vehicle scenarios in real time, predicting Time-to-Collision and Time-to-Intersection to flag dangerous situations early.

PythonTransformersKerasHuggingFace
02

Scalable Multi-client Backend with VLM

Dynamic Kubernetes pod spawning and teardown per user connection, serving multiple clients concurrently with a Vision Language Model at its core.

Node.jsPythonKubernetesDockerNginxVLM
03

CI/CD Monitoring Dashboard

Desktop dashboard with live hardware simulation data streamed via Raspberry Pi, FastAPI backend, PySide6 frontend, and Jenkins pipeline integration.

PythonFastAPIPySide6JenkinsRaspberry Pi

Let's Build
Something.

Open to roles in AI infrastructure, backend systems, and applied ML. Reach out.