Help & Supportcareers@gravityer.com
AI Engineer with less than a year in Deep Learning & Generative AI
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
I am an undergraduate researcher working in deep learning systems, with a focus on generative models and embodied AI. My work often involves close reading of primary research literature and translating theoretical foundations into working implementations - spanning diffusion models, transformer architectures, and vision-language systems. I am seeking to contribute to and grow within a foundational research environment, working on problems at the intersection of machine learning and its real-world applications.
University Institute of Technology, The University of Burdwan.
Bachelor of Engineering (B.E.) · Computer Science and Engineering (CSE)
N/A – June 30, 2027
Behavioral Cloning for Self-driving
March 1, 2026 – June 1, 2026
Achieved 90% track completion on the Udacity self-driving simulator by implementing a behavioral cloning pipeline with a fine-tuned ImageNet-pretrained model, resolving a left-turn bias in training data through targeted clockwise data collection. The remaining 10% leaves room for temporal understanding using auto-regressive models, and complete scene understanding by using all the 3 cameras in the car.
View ProjectDiffusion Simulator
March 1, 2026 – June 1, 2026
Built and deployed a conditional diffusion model that successfully generates recognizable 2D shapes (point clouds) from text prompts, using a lightweight MLP-based architecture, by implementing DDPM from scratch with BERT-generated embedding conditioning, and iterating on a key architectural insight, i.e. denoising the entire point cloud rather than individual points.
View ProjectTransformer from scratch
April 1, 2025 – June 1, 2026
Implemented a transformer encoder from scratch to learn about and understand the attention mechanism at the architectural level – positional encodings, scaled dot-product attention, residual connections, the works – deliberately without relying on any high-level abstractions.
View ProjectLatent-CLIP
April 1, 2025 – June 1, 2026
Implemented a CLIP-based Visual Question Answering model from scratch in PyTorch. Trained it on the EasyVQA dataset, and independently deployed it as a live Streamlit web app after owning the full pipeline end-to-end.
View ProjectWinner, HackWiz 2.0 Hackathon
Unknown
June 1, 2026 – Present
Published paper, "Modalities Got Latent" in Ideathon 2025, college symposium
Ideathon 2025
January 1, 2025 – Present
Finalist, Smart India Hackathon 2024
Unknown
January 1, 2024 – Present
Advanced Scholarship recipient (2023)
AWS
January 1, 2023 – Present
AWS AI/ML Scholar (2022)
AWS
January 1, 2022 – Present
Cultural Fit Analysis
The candidate's academic background, focus on foundational research, and active participation in hackathons and scholarly activities suggest a strong cultural fit for a research-oriented AI engineering role. Their interest in generative models, embodied AI, and real-world applications aligns with an innovative and impact-driven environment. The diversity of their academic projects indicates adaptability and a broad interest in various AI sub-fields.
Soft Skills & Operational Fit
The candidate demonstrates strong initiative and problem-solving skills through their project work, such as resolving left-turn bias in self-driving data and iterating on architectural insights for diffusion models. Their participation in hackathons and paper publication suggests a collaborative and innovative mindset. The ability to deploy projects independently indicates a good understanding of the full development lifecycle.