Portfolio Hub · Vishwajeet Adkine

AI Builder
& Researcher

Cross-platform AI systems, interpretable ML research, and educational platforms — engineered end-to-end from first principles.

Select a portfolio

Two portfolios live here — choose the lens most relevant to you.

Scroll
Portfolio 01
AI/ML Projects & Systems

AI/ML Projects Ecosystem

15-agent Stremini workspace, KhetiAI for Indian farmers, Electron desktop automation, Android AI keyboard, Stremini Academy with 517 topics, and deployed ML pipelines.

Cloudflare Workers K2-Think-v2 Kotlin Flutter PyTorch
Portfolio 02
Research Portfolio

Research & AI Safety Work

6 case studies & papers: AI safety, prompt injection defence, scam detection, spectral transformer theory, AgriTech field research, and two interactive ML papers with live figures.

AI Safety Transformer Theory Interpretable ML Interactive Papers

About

Built from
first principles

Every project is engineered end-to-end — mobile apps, browser automation, multi-agent platforms, and theoretical ML research — with a preference for self-healing, robust systems that handle model output unpredictability gracefully.

BSc AI applicant · Southampton. Interests in AI safety engineering, interpretable ML, and building AI ventures for underserved communities in South Asia.

PyTorch HuggingFace Cloudflare Workers Kotlin Flutter Electron K2-Think-v2 Supabase

Portfolio Overview

15+
AI Agents Built

Stremini multi-agent workspace spanning Intelligence, Creation, Strategy, and Execution layers — all deployed on Cloudflare Workers.

94.96%
Scam Detection Acc.

Interpretable ML pipeline on 28,000 URLs with 100% recall, SHAP-analogous attribution, and a live Cloudflare Worker inference endpoint.

517
Academy Topics

Stremini Academy — 25 tracks, 6 tiers from foundations through capstone, with XP tracking, quizzes, and full-text search.

University of Southampton
BSc Artificial Intelligence Applicant · 2025
Stanford University
Machine Learning Certification
DeepLearning.AI
Deep Learning Specialization
HuggingFace Hub
Published: DistilBERT IT Support Classifier

Get in Touch

Collaborate on
what matters

Open to ML research collaborations, AI safety discussions, internship opportunities, and conversations about building technology for underserved communities.

Research Interests

Spectral theory of neural networks · AI safety & prompt injection defence · Interpretable ML for security · Multi-agent architectures · Agricultural AI for South Asia

Models in Production

MBZUAI K2-Think-v2 · Gemini 2.5 Flash · Groq Llama 4 Scout · DistilBERT · Groq Llama 3.3-70b

Languages

Python · JavaScript / TypeScript · Kotlin · Dart (Flutter)

Collaboration Areas

AI for social impact in South Asia · Educational AI platforms · AI safety tooling · Applied ML research with deployable outcomes