Skip to content
View katgharalad's full-sized avatar
5 followers · 7 following
  • Columbus, Ohio
  • 18:28 (UTC -04:00)

Achievements

Achievement: QuickdrawAchievement: Pull SharkAchievement: YOLO

Achievements

Achievement: QuickdrawAchievement: Pull SharkAchievement: YOLO

Block or report katgharalad

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
katgharalad/README.md

Aarav Singh

Building systems that turn unstructured data into decision pipelines.

Work

OPEF – Phase I ESA Automation

End-to-end system for automating environmental due diligence workflows for Phase I Environmental Site Assessments.

  • Parses regulatory and environmental datasets from EPA, state records, and geospatial sources
  • Builds an evidence graph to connect: raw record → extracted fact → canonical entity → spatial relation → risk classification
  • Automates REC (Recognized Environmental Condition) identification
  • Integrates document analysis, geospatial screening, and structured reasoning into one pipeline

Impact
$3000 / 2 weeks → ~$500 / <10 hours

Technical Stack

  • Backend: Python, FastAPI
  • Cloud / Infra: AWS S3, AWS Lambda, PostgreSQL
  • Geospatial: Google Earth Engine, Sentinel-2, GEDI, EPA APIs
  • Extraction / NLP: embeddings, rule-based extraction, document parsing
View technical flow 
flowchart LR
    A[Raw Regulatory / Environmental Data] --> B[Parsing + Extraction]
    B --> C[Extracted Facts]
    C --> D[Canonical Entity Resolution]
    D --> E[Spatial Relationship Engine]
    E --> F[Risk Classification]
    F --> G[REC / ESA Finding]
Loading

Mentor Connector

Matching system for founders ↔ mentors using hybrid retrieval.

  • Combines dense embeddings (semantic similarity) + BM25 (lexical ranking)
  • Ranking and filtering pipeline for high-precision matching
  • Handles both structured and unstructured profile data
  • Designed to avoid noisy keyword-only matches

Technical Stack

  • Python
  • Vector embeddings
  • BM25 (information retrieval)
  • REST APIs

TSheets Processing System

Operational ETL pipeline for time and work logs.

  • Ingests raw logs → cleans → normalizes → structures
  • Handles inconsistent schemas and noisy inputs
  • Outputs analytics-ready datasets for downstream workflows

Technical Stack

  • Python
  • SQL
  • ETL pipelines

LateGrub

Full-stack food ordering platform.

  • Real-time order flow and backend handling
  • Authentication, session management, database integration
  • Built with focus on system design and execution

Technical Stack

  • JavaScript
  • Backend APIs
  • Database systems

OPEF Website

Product-facing system for technical and non-technical users.

  • Communicates environmental analysis workflows clearly
  • Built for demos, pilots, and stakeholder onboarding
  • Structured for clarity and conversion

Technical Stack

  • React
  • Frontend systems
  • Deployment pipelines

Technical Scope

Languages

Python C++ SQL R Scheme JavaScript

Machine Learning

Scikit-learn TensorFlow Computer Vision Geospatial ML

Infra / Systems

AWS (S3, Lambda) Azure Linux Scripting and automation ETL pipelines REST APIs

Geospatial / APIs

Google Earth Engine NASA Earthdata Copernicus Sentinel GEDI EPA datasets

Focus

  • Systems over scripts
  • Real data over toy datasets
  • Speed and execution
  • Traceable, production pipelines

Contact

aarav@opef.ai

Pinned Loading

  1. lategrub lategrub Public

    TypeScript

  2. audiotrancriptionfire audiotrancriptionfire Public

    Python 1