Skip to content
pushkal.

Projects

Things I've built

Research systems, open-source tooling, and products. The featured set is where most of my time goes; expand to see everything.

VALOR-RAG

Under Review — HICSS 2026

Value-aware security framework for enterprise RAG

A security framework for retrieval-augmented generation systems focused on improving trust, reliability, and robustness against retrieval failures and adversarial manipulation. Incorporates trust signals and retrieval quality analysis for enterprise RAG deployments. Under review at HICSS 2026.

  • Trust-aware retrieval analysis over enterprise knowledge bases
  • Context quality evaluation for generated responses
  • Retrieval security mechanisms against adversarial manipulation
  • Adaptive defense strategies tuned to enterprise RAG pipelines

Architecture

  1. Retrieval

    Vector store query with candidate document set

  2. Trust Scoring

    Trust-signal analysis over provenance and retrieval quality

  3. Context Validation

    Quality evaluation of retrieved context before generation

  4. Adaptive Gate

    Defense strategy that admits or flags low-trust context

  5. Generation

    LLM generates only from validated, trust-weighted context

Python · PyTorch · Transformers · Vector Databases

Meridian

Founder

AI-native CRM for tutoring organizations

An AI-native CRM platform designed for tutoring organizations and education businesses. Generic CRMs model deals; tutoring businesses run on recurring sessions, multi-tenant organizations, and student-level workflows — Meridian models that directly.

  • Multi-tenant organization management with per-org data isolation
  • Student and session workflows built for tutoring businesses
  • Scheduling infrastructure for recurring sessions
  • Analytics and operational tooling for education businesses

Next.js · TypeScript · FastAPI · PostgreSQL

RAGuard

Open-source RAG security toolkit

An open-source toolkit for protecting RAG applications against retrieval attacks and prompt injection threats. Ships as a drop-in middleware layer: analyze retrieval context and detect adversarial patterns before they reach the model.

  • Retrieval context analysis before generation
  • Attack detection mechanisms for common RAG threats
  • Security middleware that drops into existing RAG pipelines

Python · FastAPI · Transformers

Context Bridge

Persistent memory for AI coding agents

A memory and checkpointing system for AI coding agents that preserves goals, decisions, blockers, and progress across sessions so an agent resumes exactly where it left off instead of rediscovering state.

  • Session checkpointing with goals, decisions, and blockers
  • Cross-session context restoration on startup
  • Progress tracking to prevent repeated failures across sessions

Python · FastAPI · SQLite

Pulse

Mobile app for organization management

A mobile application built for organization management and member engagement — member profiles, scheduling, resources, and community features in one app.

  • Member profiles and directory
  • Calendar systems for organization events
  • Shared resources and document access
  • News and community engagement features

Flutter · Firebase · Riverpod

LitMind

AI-powered research assistant

An AI-powered research assistant Chrome extension for improving literature review workflows — semantic annotation, research organization, and knowledge discovery in one place.

  • Semantic annotation directly over paper text
  • Research organization tools for literature review
  • Knowledge discovery from prior reading

TypeScript · React · Embeddings

Margined

LLM unit economics platform

An LLM unit economics platform that helps AI companies understand inference costs, margins, and profitability. Connects inference expenses to features and customers so teams can see which usage is profitable and which is subsidized.

  • Per-feature and per-customer inference cost attribution
  • Margin and profitability dashboards

TypeScript · Next.js · PostgreSQL