// builder · engineer · founder

Aryaman Singh Rana

I build AI systems,
data pipelines, and products people use.

02

Experience

Self Employed

Indie iOS App Developer

Remote

  • Independently built and shipped two full-stack iOS app using React Native and TypeScript, with FastAPI (Python) backend hosted on Azure AppService - covering system architecture, onboarding flows, feature implementation and third party integrations.
  • Architected a multi-provider AI pipeline (Groq, Azure OpenAI) with graceful fallback handling, powering transcription, sentiment analysis and affirmation generation across 4 distinct AI workflows.
  • Conducted competitive gap analysis across multiple iOS apps, mapping UI/UX patterns in Figma and validating market demand through social media before writing a single line of code.
  • First app took 2mon to build and get approved on AppStore. Second app's V1 was shipped in 1week with Zero AppStore rejections.
  • Drove 110K+ organic views across Instagram (80K), YouTube (30K), and TikTok in 3 weeks by producing 55+ original short-form videos, creating AI-generated avatars for scaled content, and managing 5 accounts with 2-3 posts per day.
React Native Typescript JavaScript Python FastAPI Azure iOS Development Marketing Product Development PostgreSQL

Artha Intelligence

Co-Founder and Software Developer

Bengaluru, India

  • Multi Agent System for IPO Assesment:
  • Single-handedly built a LangGraph-based framework to analyze IPO-bound companies, integrating Red Herring Prospectus data and 200+ articles per report, targeting 30% analysts time savings.
  • Deployed 15 agents across 5 sectors, using sentiment analysis and dynamic question formulation to increase report accuracy by 20%.
  • Independently designed and implemented a RAG system to extract insights from RHP, improving retrieval accuracy from 66% to 81% through iterative prompt engineering and retrieval optimization.
  • Contextual RAG:
  • Built a Contextual RAG - a document processing pipeline using Azure services, subquery generation, hybrid search (BM25 + vector search), query fusion and respective storage in Qdrant, preserving contextual integrity of financial documents.
  • Was selected for AltSpace AI Launchpad - 20+ developers from across India were sponsored to stay in Dharamshala and work on their respective start-ups.
Python RAG Contextual RAG Generative AI LangGraph LangChain ReactJS Vector Databases supabase Qdrant

Tekion Corp

Associate Software Engineer

Bengaluru, India

  • Documented and analyzed logic for 100+ pre-existing patterns, then developed custom logic for 8 file formats (CSV, DAT, XLSX, XLSB, XML, TXT, JSON, PDF), consolidating said patterns, removing code debt and enhancing efficiency by 75%.
  • Automated 40% of manual processes with scripts for creating, updating and validating raw schemas for 49 types.
  • Implemented performance-based logging and status tracking, segmenting by execution metrics and improving system monitoring coverage by 60%.
  • Coded and optimized model developments and collaborated with testing teams to resolve intricate model logic, investing extensive hours to ensure accuracy.
  • Benchmarked PySpark performance in AWS Glue across multiple partitions - measuring execution times, DataFrame write speeds, and Athena query latency across datasets ranging from thousands to millions of rows.
  • Tested single-threaded vs. multi-threaded configurations, identifying optimal concurrency settings that informed production deployment decisions.
Pyspark Python AWS AWS DynamoDB AWS Glue AWS EC2

Tekion Corp

Associate Software Engineer Intern

Bengaluru, India

  • Learnt PySpark and coded business logic, with the aim of optimizing data workflows and ensuring scalability for large datasets.
  • Spearheaded the conversion of UDF functions to base PySpark with the goal of reducing execution time.
  • Coded multiple automation scripts working with AppleScript, Python, AWS S3 and Rest APIs.
AWS AWS EC2 AWS DynamoDB AWS Glue Python PySpark Sql

AB InBev India

Web Development Intern

Remote

  • Initiated the ideation and creation of a POC, which aims to improve communication between the desired parties by 50%.
  • Assisted in development of the frontend of a web application which facilitates query posting and aims to accelerate problem resolution by 45%.
HTML / CSS / JavaScript ReactJS JavaScript Python TailwindCSS CSS Git
03

Skills

Technical

Python PySpark Langchain LangGraph CrewAI SQL C++ JavaScript React Native TypeScript HTML/CSS

Tools

AWS Glue EC2 S3 Athena DynamoDB Lambda Azure AppService Azure Functions Supabase Git ComfyUI LoRA (SDXL + Flux)
04

Projects

Flirting Drills AI dating coach iOS app — screen 1 Flirting Drills AI dating coach iOS app — screen 2 Flirting Drills AI dating coach iOS app — screen 3

iOS App |

Flirting Drills

Gamified AI dating coach with a weighted scoring engine grading user responses across multiple social skill dimensions. Full RevenueCat subscription flow with paywall gating and Supabase-backed per-user skill tracking. 3.08% App Store conversion rate.

React Native FastAPI LLM RevenueCat Supabase
→ View Project
Rewire SheEcho AI wellness iOS app — voice journaling screen 1 Rewire SheEcho AI wellness iOS app — affirmations screen 2 Rewire SheEcho AI wellness iOS app — CBT screen 3

iOS App |

Rewire / SheEcho

AI-powered mental wellness iOS app with voice journaling and CBT-based affirmations. Built a multi-provider AI pipeline (Groq, Azure OpenAI) powering transcription, sentiment extraction, and affirmation generation across 4 distinct workflows. Supabase backend with row-level security and personalized AI context per user session.

React Native TypeScript FastAPI Azure Supabase Whisper
→ View Project
Artha Intelligence IPO multi-agent LangGraph platform — AI financial analysis

AI Platform |

Artha Intelligence

Multi-agent LangGraph system for IPO assessment — 15 agents across 5 sectors, integrating 200+ articles per report. Paired with a Contextual RAG system that improved retrieval accuracy from 66% → 81% via hybrid search and query fusion. 55+ sign-ups in first month.
We've shut this project down.

LangGraph Python Azure Qdrant RAG
→ View Project
Financial Partner Chrome extension GPT-4o financial analysis — screen 1 Financial Partner Chrome extension GPT-4o financial analysis — screen 2 Financial Partner Chrome extension GPT-4o financial analysis — screen 3 Financial Partner Chrome extension GPT-4o financial analysis — screen 4

Chrome Extension |

Financial Partner

Chrome extension for real-time financial text comparison across 3 risk personas — conservative, moderate, aggressive. Sub-5-second responses for 90% of queries. GPT-4o chatbot with 85% query resolution accuracy. 500+ active users, 2,000+ total downloads.
We've shut this project down.

JavaScript Azure Functions GPT-4o Supabase
→ View Project
05

Side Projects & Experiments

AI financial analysis agent ecosystem view — RAG knowledge graphs AI financial analysis agent investment recommendation output

AI |

Financial Analysis Agent

An AI agent that analyzes corporate announcements and market data to provide investment recommendations with reasoning. Uses RAG, Knowledge Graphs, Hybrid-RAG, KAG, and KG-enhanced agents.

RAG Knowledge Graphs Hybrid-RAG Python
RAG implementation demo — LangChain AWS Aurora Llama 3 question answering

AI |

RAG Implementation

A question-answering system that lets users point to data sources — codebases, PDFs, documentation — and get contextual LLM responses. Built with LangChain, AWS Aurora, and Llama 3.

LangChain AWS Aurora Llama 3 Python
→ GitHub

AI |

Cold Email Generator / EmailGenie

AI-powered email copywriting tool for personalized outreach to businesses, job seekers, and freelancers. Generates context-aware emails at scale.

Groq APIs Streamlit LangChain Python
DeLand Image

Blockchain |

Decentralised Land Management / DeLand

A decentralized system for storing national land records on-chain, applying data decentralization concepts to India's land registry problem.

React JavaScript Moralis Solidity
→ GitHub
Personal Chrome new tab extension — Memento Mori Bhagavad Gita reminder

Personal |

New Tab Extension

Every tab I open should remind me of "Memento Mori" and the Bhagavad Gita — to keep me focused on limited time and impact. Code available upon request.

Browser Extension JavaScript
Personal minimal todo app — cyberpunk design Netlify

Personal |

Todo App

I regularly build tools to automate parts of my life. This is a simple todo app which I use to manage my day to day life. As constraints breed creativity, I knowingly restricted a lot of features and kept it as simple as possible.

JavaScript CSS
→ View
Mooood of Elon LoRA composite — GTA V style Elon Musk comic strip Mooood of Elon LoRA GTA V Elon Musk — frame 1 Mooood of Elon LoRA GTA V Elon Musk — frame 2 Mooood of Elon LoRA GTA V Elon Musk — frame 3

Creative |

Mooooood of Elon

Combined multiple LoRAs — a GTA V style model and an Elon Musk face model — to create comic strip memes with a distinctive aesthetic.

Diffusion Models LoRA ComfyUI
GTA V LoRA style transfer diffusion model — sample 1 GTA V LoRA style transfer diffusion model — sample 2 GTA V LoRA style transfer diffusion model — sample 3

Creative |

GTA V Styled LoRA

My brother asked if I could replicate the GTA V art style in AI-generated images, which led me to experiment with LoRAs and train my own GTA V style model from scratch.

Diffusion Models LoRA SDXL ComfyUI
DreamBooth fine-tuned diffusion model Tokyo Revengers Mikey — sample 1 DreamBooth fine-tuned diffusion model Tokyo Revengers Mikey — sample 2 DreamBooth fine-tuned diffusion model Tokyo Revengers Mikey — sample 3 DreamBooth fine-tuned diffusion model Tokyo Revengers Mikey — sample 4 DreamBooth fine-tuned diffusion model Tokyo Revengers Mikey — sample 5

Creative |

DreamBooth Training

Fine-tuned a diffusion model using character images from Tokyo Revengers to generate similar-featured characters in the same visual style.

DreamBooth Diffusion Models Python
05

Education

Shiv Nadar University

BTech in Computer Science & Engineering

Greater Noida, UP  ·  Graduated August 2023