AI Systems |

AI Market Intelligence Terminal

Sun, 01 Feb 2026 00:00:00 +0000

A solo-built market intelligence platform targeting commercial deployment across Indian equities, global crypto, US markets, and commodities, priced in USD and INR.

Architecture

Layer 1 - Data aggregation. Live market data from CoinGecko, Alpha Vantage, NSE Direct, and GoldAPI flows through n8n using a cascading merge strategy designed around n8n’s 10-input node limit.

Layer 2 - Parsing and delivery. An LLM-based parser via OpenRouter applies strict anti-hallucination guardrails, then delivers daily reports as formatted HTML email and Discord embeds through a custom markdown-to-HTML renderer.

Layer 3 - The LLM Council (live). GPT-4o-mini, Gemini Flash, and Claude Sonnet act as independent analysts, with Claude Opus as chairman synthesising a final report through anonymised peer review. The FastAPI backend is containerised on Railway and called from n8n via authenticated webhooks.

In progress. Layer 4: RAG pipeline on Pinecone. Layer 5: conversational Next.js frontend with OpenAI Whisper STT/TTS, and multi-user scaling on Supabase.

Flux - AI Goal Management

Thu, 29 Jan 2026 00:00:00 +0000

Flux is an AI-powered goal and task management application built as the capstone of the Outskill AI Engineering Fellowship (Cohort 3, Team 8), where I served as Technical Lead.

Technical contributions

I designed the monorepo architecture and implemented three agents from scratch: the GoalPlannerAgent, Scheduler Agent, and Classifier Agent, orchestrated with LangGraph. The RAG pipeline runs on Pinecone with text-embedding-3-small, and source citations are surfaced directly in the UI.

Debugging under deadline

Drove all 24 integration tests to green by diagnosing JWT HS256/ES256 mismatches, auth sync gaps, and route contract divergences. Resolved a critical silent failure where VITE_USE_MOCK=true masked every backend call. Managed the team’s Git branch strategy using archival tags to preserve contribution history.

PulseFit - AI Fitness Tracker

Sat, 01 Nov 2025 00:00:00 +0000

An AI-powered mobile fitness application combining workout tracking, nutrition analysis, and personalized coaching.

Computer vision handles meal analysis, an LLM generates workout recommendations, and real-time activity monitoring feeds progress analytics. Built on a React Native cross-platform architecture with a Supabase backend, OpenRouter AI APIs, edge functions, and row-level security, delivering one experience across iOS, Android, and web.