Shivam Padmani

Senior Data Scientist · AI/ML Researcher
6+Projects
1Paper
8K+Reach
01

Selected Projects

NLP · RAG 01
Vachanamrut Guide

Conversational AI system over unstructured spiritual text. Converted 10K+ lines into an optimised SQL database, preserving semantic integrity. Deployed as a browser-based agent with on-edge processing for data privacy.

~90% response accuracy · 40% API cost reduction
RAGSQLBrowser AgentLLM
Gen AI · NLP 02
Poetry Generation in Low-Resource Indic Language

Fine-tuned a seq2seq translation model on 8000 Gujarati couplets of Bhaktchintamani to generate poetry in the style of Poet Nishkulanand Swami from English prompts. Deployed as a public web application.

8000+ people reached in the first week of launch
Seq2SeqGujarati NLPFine-tuningWeb App
Computer Vision 03
Image Generation from Scratch

Implemented GANs, VAEs, and Diffusion Models entirely in raw PyTorch, including DCGAN, conditional GAN, beta-VAE, VAE-GAN, DDPM, i-JEPA, Score Matching, and Implicit Sampling directly from research papers.

FID < 100 on denoised diffusion probabilistic models
PyTorchDiffusionGANVAE
NLP · Geospatial 04
Geolocation Prediction from Twitter User Data

Led a team of 3 in a Kaggle-based competition to predict user geographical coordinates from tweet data. Deployed a fine-tuned model mapping user sentiments to specific geographical coordinates via NLP.

RMSE of 10.3 on geolocation coordinates
NLPSentiment AnalysisFine-tuningKaggle
02

Skills & Tools

Machine Learning
  • PyTorch
  • GANs, VAEs, Diffusion Models
  • LLM Fine-tuning
  • Kolmogorov-Arnold Networks
  • Seq2Seq / Transformers
  • Scikit-learn / XGBoost
Agentic AI
  • Agent Federation
  • RAG Pipelines
  • Knowledge Graphs
  • Agno Framework
  • Conversational AI
Data Engineering
  • PySpark
  • SQL
  • Google Cloud Platform
  • OpenFOAM (HPC)
  • Image Processing
Languages
  • Python
  • R
  • MATLAB
  • SQL
  • LaTeX
Statistics
  • Bayesian Learning
  • Probabilistic ML
  • Random Forest / Ensemble
  • Clustering (k-means)
  • CFD & Physics-ML
Other
  • Linux / HPC environments
  • Virtual Labs development
  • Network Traffic Estimation
  • Fluid Simulation (Re<1000)
03

Publications & Research

001 2025
Function Regression using Forward-Forward Training and Inferring Paradigm
Padmani S., Joshi A. · MTech Research Project · Indian Institute of Science
Under Review

Proposed a novel function regression approach using neural networks that avoids backpropagation, exploring new learning mechanisms via unidirectional (forward-forward) training. Implemented Kolmogorov-Arnold Networks (KAN) to reduce parameter count while achieving MSE of 0.001.

04

Work Experience

Jul 2025 — Present
Mphasis
Current
Senior Data Scientist

Delivering client-facing demos and whitepapers in applied AI. Working across network intelligence, knowledge management, and autonomous agent systems.

Outperformed SOTA in network traffic matrix estimation for client Built knowledge graph over procedural docs for LLM Q&A with reduced hallucinations Automated CRO alert resolution via AI agent federation on AWS
Jan 2025 — May 2025
Nosh Robotics
AI Intern

Developed AI features for a food robotics product — from recipe scaling to agentic cooking command generation from custom inputs.

Predicted recipe patterns and built multi-serving scaling algorithm Built agentic pipeline for machine-readable cooking commands using Agno Classified recipes via k-means clustering on consistency & ingredient quantities
May 2022 — Jun 2022
IIT Bombay · CFD Lab
Research Intern — Prof. Atul Sharma

Applied machine learning to computational fluid dynamics, predicting intermediate timesteps and reducing simulation overhead for laminar flows.

Reduced simulation time by 40% using AI-assisted CFD for Re<1000 flows Reduced net computational cost by 10% via physics-ML integration Used OpenFOAM on HPC Linux environments to solve fluid mechanics problems
05

Education

2023 — 2025
MTech — Indian Institute of Science
Bengaluru, India
GPA: 8.50 / 10. Research focus on novel neural network training paradigms and function regression without backpropagation. Led campus initiative for inclusive food access for 80+ students.
2019 — 2023
BTech — Nirma University
Ahmedabad, India
GPA: 8.94 / 10. Strong foundation in algorithms, image processing, and applied ML. Final year project on screw-nut detection using image processing techniques (90%+ accuracy).
06

Beyond Work

Achievement
Silver Developer recognition for creating Virtual Labs on Helmholtz Resonator and Hologram.
Leadership
Led campus initiative at IISc to provide onion-garlic free meals, impacting 80+ students' daily food access.
Social Activity
3 years of National Cadet Corps (NCC) Army Training, earning "Alpha" grade in all written and practical examinations.
Volunteering
Content Manager and Writer for Bal Prakash — a value-based English kids magazine by the BAPS Organisation.