Help & Supportcareers@gravityer.com
Make AI fast again
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Spleenlab GmbH
Data Scientist
June 27, 2026 – Present
YOLO-World-TensorRT-CPP
October 5, 2025 – October 5, 2025
A high-performance, real-time open-vocabulary object detector using YOLO-World, C++, and TensorRT.
View Projecttriton-swiglu
October 4, 2025 – October 4, 2025
An implementation of a fused SwiGLU activation kernel using OpenAI Triton, designed to accelerate modern Large Language Models (LLMs) by reducing memory bandwidth bottlenecks.
View Projectflash_attention
October 4, 2025 – October 4, 2025
This repository contains a self-contained, from-scratch implementation of the FlashAttention algorithm in a single CUDA C++ file. The goal is to provide a clear, raw, and focused demonstration of the core concepts that make FlashAttention I/O-aware and significantly faster than standard attention implementations.
View Projectblog_code
May 25, 2025 – May 27, 2025
This is a repository to store code associated with my blog series on High Performance Computing https://sanket-pixel.github.io/blog/
View Projecttensorrt_cpp
August 7, 2023 – August 28, 2023
Contains code for performing TensorRT inference using C++ on Image Classification Resnet.
View Projecttensorrt_deeploy
May 7, 2023 – July 24, 2023
This repository contains source code for my blog "TensorRT and it's ways".
View Projectstereo-depth
August 14, 2021 – September 29, 2021
Estimating distance and depth using information from left and right cameras
View Projectattentive-mot
August 12, 2021 – September 6, 2021
Multi-Object Tracking using Transformer Networks with Soft Data Assosiation
View Projectaction-recognition
June 10, 2021 – August 10, 2022
This repository contains my projects related to video analytics with topics like Video clip classification, Temporal video segmentation, Spatio-temporal action detection, Saptio-temporal modeling of humans and objects, Anticipation, Weakly supervised learning, Video summarization, Affordance, Egocentric videos, Multi-person tracking .
View ProjectCultural Fit Analysis
The candidate's projects show a strong inclination towards cutting-edge research and performance optimization in AI/ML, which aligns well with an innovative and technically driven culture. The personal projects demonstrate initiative and a passion for the field. However, the lack of team-based projects or collaborative descriptions makes it difficult to fully assess cultural fit beyond technical alignment.
Soft Skills & Operational Fit
Insufficient data to assess soft skills and operational fit. The candidate's project descriptions indicate a strong technical focus and ability to work on complex, self-directed projects.