AI-native mini CRM backend for reaching D2C shoppers. Two FastAPI services that share nothing at runtime and communicate over HTTP only:
crm_api/ is the main CRM: ingest, segments, campaigns, AI endpoints, and the receipt callback API.
channel_service/ is a stubbed messaging channel: it simulates delivery outcomes and calls back into the CRM receipt API.
The product bet is explainable AI: every AI output ships with the reasoning that produced it, so a human can verify the decision before acting on it.
2
41
The AI/ML roadmap everyone shares is either overwhelming or outdated.
So I wrote the guide I wish I had when I started.No Gate, No Problem is a practical AI/ML career guide for students from tier-2/3 colleges, covering what to learn, what to skip, how to build real projects, and how to land opportunities without an IIT tag. This is just the beginning.
It's my first digital product, and I'd genuinely love your feedback. ❤️
Get it here: https://contra.com/products/y88ymoKM-no-gate-no-problem
1
40
SheSpeaks is one of the flagship initiatives I have worked on under the ABES ACM-W Student Chapter. The initiative was designed to create a space where women could openly share their journeys, challenges, ideas, and experiences around technology, education, leadership, and personal growth. I contributed to the planning, coordination, outreach, and execution of the initiative while working closely with the team to ensure meaningful participation and engagement. What made SheSpeaks special to me was that it was not just a technical event, but a platform focused on confidence, visibility, communication, and community-building for women in tech. Working on it helped me understand the importance of creating supportive ecosystems alongside technical growth.
1
38
XAI Forensics: Explainable Sentiment Analysis Tool
Built a full-stack Explainable AI platform to explore how transformer models interpret and classify sentiment. The system compares predictions from DistilBERT and RoBERTa using multiple explainability techniques including LIME token attribution, counterfactual perturbation, and dual-model disagreement analysis.
Implemented a FastAPI backend with latency tracking and deployed the Dockerized inference pipeline on HuggingFace Spaces, with a forensic-style interactive dashboard built using Next.js and deployed on Vercel.
Key highlights:
• Compared 2 transformer models trained on different domains
• Implemented 300-sample LIME explanations with top-token attribution
• Built 4 REST APIs for analysis workflows
• Added counterfactual sentiment-flip analysis and disagreement scoring
• Deployed end-to-end at zero hosting cost using Vercel + HuggingFace
Tech Stack: Python, FastAPI, HuggingFace Transformers, LIME, Docker, Next.js
GitHub: https://github.com/parshvi1508/XAI_Forensic
(https://github.com/parshvi1508/XAI_Forensic)Demo: https://xai-forensic.vercel.app/
1
90