Faiz Khan

AI & Backend Developer

Computer Engineering student at TSEC building production-grade RAG systems, and scalable backend services. Skilled in Python, C/C++, and Go, with a focus on performance, reliability, and clean architecture.

$neofetch
OS: Backend & AI Engineering
Kernel: RAG Systems · AIML Infrastructure
Uptime: ~2 years active development
Shell: Python, C/C++
Current Task: Scaling RAG pipelines and low-latency trading systems
$_

01. System Status

Computer Engineering student at TSEC focused on AI systems and backend infrastructure. Currently working as an SDE Intern building low-latency trading infrastructure, event-driven APIs, and broker abstraction layers for production deployments.

Experienced in designing production-grade RAG pipelines, multi-agent evaluation systems, and cloud-native AI services. Winner of multiple AI hackathons (DOC.AI, Codeissance, Techathon 2025) with projects spanning governance AI, multimodal audio intelligence, and scalable backend architectures.

Python (FastAPI, Flask, Streamlit)
RAG & AI Systems (Embeddings, FAISS, Multi-Agent)
Backend Systems (APIs, WebSockets, Event Flows)
Databases (PostgreSQL, MySQL, Redis, FAISS)
C/C++

02. Experience

Software Development Engineer Intern

Nidhi-Harsh Wealth Ventures

Jan 2026 - Feb 2026
Remote
  • Contributed to AlgoBridge, adding WebSocket-driven price triggers and event-based order flows to keep NIFTY/BANKNIFTY executions reliable under sub-second constraints.
  • Built an abstraction layer for FYERS orders and webhooks so strategies deploy once across brokers.
  • Delivered a Flask REST API and Streamlit dashboard so non-technical operators can configure strategies.
  • Implemented an option instrument resolver caching 6,000+ symbols to translate human-readable inputs into broker-specific instruments, reducing lookup latency and order errors.

03. Tech Stack

Languages

PythonCC++SQL

Databases

MySQLPostgreSQLSQLiteRedisFAISS (Vector DB)

Frameworks & Libraries

FastAPIFlaskReactStreamlitWebSockets

AI/ML & NLP

EmbeddingsRAGFAISS indexingMulti-agent

Cloud & DevOps

AWS (S3, Lambda, Bedrock)GCP Cloud Run

Tools

n8nLinuxGitVS CodePostman

04. Deployed Protocols

Nolan – AI-Powered Screenplay Generator & Analyzer

Local LLMs • Multi-Agent AI Systems • Knowledge Graphs • FastAPI

  • AI-driven IDE for writers with context-aware assistance and privacy-first local LLM execution, bridging native and web environments seamlessly.
  • Multi-agent LLM Council with Director/Editor/Audience roles maintains narrative consistency, identifies plot holes, and maps idea relationships via knowledge graphs.
  • FastAPI-deployed privacy-focused system won 1st Place at TSEC-HACKS 2026, delivering advanced AI evaluation for serious authors.

AETHER – AI Governance & Adjudication Engine

LLM Council, RoBERTa, Multi-Agent Debate

  • Designed AETHER as a modular AI governance and adjudication engine using multi-agent debate with automated judging to deliver consistent rulings.
  • Implemented pluggable evaluator modules so domain teams can drop in specialized reasoning.
  • Added traceable reasoning graphs and decision logs to support transparency, post-hoc analysis, and regulator-ready evidence.
> GitHub_Repo

Mind Voice – Depression Detection System

Agentic AI, ML, Audio Processing, GCP

  • Designed a multimodal classification pipeline mixing acoustic signals with linguistic sentiment and keyword cues to improve robustness across varied speakers.
  • Integrated Speech-to-Text with emotion scoring so text and acoustic signals reinforce each other, improving prediction reliability across accents and recording quality.
  • Deployed on GCP Cloud Run behind API Gateway, enabling fast iteration without redeploying clients.
> GitHub_Repo

DoChat – AI Document Assistant (RAG System)

FastAPI, FAISS, React, TypeScript, Groq LLM, CrossEncoder

  • Full-stack RAG application enabling conversational Q&A over PDF documents with FastAPI backend, FAISS vector indexing, and Groq LLM integration.
  • Implemented dual-stage retrieval (vector search + CrossEncoder reranking) with React + TypeScript frontend for real-time document processing.
  • Architected modular components with configurable parameters for seamless upload-to-query workflows and production-grade scalability.
> GitHub_Repo

Goodwill Copy Trader

FastAPI, Python, Excel, GWCIndia API

  • Automated trade replication system for real-time order mirroring.
  • 100% order synchronization across multiple accounts.
  • Secure token exchange & Excel-based configuration.
> GitHub_Repo

n8n Audio Processing

n8n, Flask API, FFmpeg, Docker

  • Workflow for automated audio pitch & tempo modification.
  • Uses n8n to orchestrate uploads, API calls, and response handling.
  • Audio processed via custom Flask-FFmpeg.
  • Triggers using new file upload in Drive.
> GitHub_Repo

05. Milestones

2026

Nidhi-Harsh Wealth Ventures SDE Intern

Building low-latency trading infrastructure, event-driven APIs, and broker abstraction layers for production deployments.

2026

1st Place - TSEC-HACKS 2026

24-hour AI/ML Sprint - Nolana, an AI-powered screenplay generator and analyzer.

2026

2nd Place - TSDC 2026

24-hour AI/ML Sprint - Built AETHER, an AI governance and adjudication engine.

2025

1st Place - Techathon 2025

8-hour AI/ML Sprint - Built LLM Council for AI-vs-human text classification using adversarial LLM arguments and fine-tuned RoBERTa judge.

2025

1st Place - Codeissance 2025 (TSEC)

Mind Voice - multimodal ML pipeline combining acoustic and linguistic signals for depression detection.

2024

1st Place - DOC.AI Hackathon

48-hour hackathon - Built DoChat, an advanced RAG system with FAISS indexing, contextual Q&A, citation tracking, and multi-document analytics.

2024

Certifications

CS50SQL (Harvard University) • Python in Data Science (IIT Madras NPTEL) • SQL Certified (HackerRank)

2023 - 2027

B.E in Computer Science

Thadomal Shahani Engineering College (CGPA: 8.25)

05. Establish Connection

Currently open for internships and collaborative projects involving RAG, Backend Systems, or Algo Trading.