Python backend and data engineer, Kathmandu
I build Django backends and write web scrapers, including for sites that try to block them.
I'm in the final year of my BIT at Patan Multiple Campus and looking for backend or data engineering work. Most recently I was a part-time data engineer at Ankamala Research.
GET/about
I'm Prashant, a backend developer from Kathmandu. Most of my work sits between Django on one side and web scraping on the other.
On the Django side I've written custom middleware, JWT-secured REST APIs and signal-driven features, with tests to back them up. I've also moved a project's data from SQLite to PostgreSQL.
Scraping started at Parsedom, a Toronto company, where I built scrapers for more than 20 websites. I kept at it with freelance clients and a part-time data engineering job, and most of what I know there I learned by testing what works rather than reading about it.
Outside programming I spend a lot of time on philosophy, physics and maths, and I write short essays on philosophy.
GET/experience
Specific tools, data sources and partners are covered by a non-disclosure agreement.
GET/projects
A full-stack collaborative productivity tracking platform where teams create rooms, run tracked sessions, log tasks, and compete on live leaderboards. Features a JWT-secured REST API, 3 custom middleware layers, event-driven Django signals architecture, and 120+ test cases.
$ pip install git+https://github.com/Pr45H4nt/aiscrapelite.git $ aiscrapelite "https://news.ycombinator.com" \ "get the top 5 post titles" 1. Opening https://news.ycombinator.com 2. Analyzing page... 3. Getting commands from LLM... No more commands needed - data should be visible 5. Converting page to markdown... 6. Extracting information... {"valid": true, "data": [...], "errors": []}
Give it a URL and describe the data you want in plain English, and it hands back JSON. On each step it lists what on the page can be clicked or typed into, asks an LLM what to do next, and carries that out in Playwright. Once the data is on screen it converts the page to markdown, pulls out what you asked for and checks any fields you marked as required.
# find dead links and check whether each domain still answers $ python extract_dead_links.py sample/links.json -o dead_links.json # search, fall back to the Wayback Machine, rank with an LLM $ python find_candidates.py -i dead_links.json -o candidates.json # a person picks the right replacement $ python review_candidates.py \ -i candidates.json -o final_replacements.json
Finds where a dead link went. It checks whether the page's domain still answers and searches that site for the title. When the domain is gone too, it reads the old title from the closest Wayback Machine snapshot and searches on that. An LLM scores the candidates, and a review step in the terminal leaves the final call to a person, because two pages can look alike without being the same document.
A resume builder in Django. A multi-step form takes you through personal details, education, experience, projects, skills and achievements. Then you pick one of six templates and download a PDF rendered with WeasyPrint. There's no frontend framework, only Django templates.
A Django social app for sharing opinions. People write, edit and delete their own posts, comment on other people's, and like or dislike them, with accounts for signing up and logging in.
A Scrapy spider for Social Blade's top 100 YouTube channels. It collects 21 fields per channel, from subscriber counts to income estimates, and gets past 403s and redirects with request headers and caching.
Python · Scrapy
Finds contact emails on company websites, drafts an inquiry for each with Gemini, and sends it once you've reviewed it. A log makes sure no company gets the same email twice.
Python · BeautifulSoup · Gemini
A Discord bot that talks like L from Death Note. It runs on Gemini, remembers the recent conversation, and waits a moment as if typing before it replies.
Python · discord.py · Gemini
A blogging site with comments and post management, plus a Django REST Framework API that the frontend talks to.
Django · DRF · Bootstrap
A Django REST Framework backend for a task app, with user authentication, task CRUD and team collaboration.
Django · DRF
Scrapes movie data from IMDb's sci-fi genre listings with Scrapy.
Python · Scrapy
Works through business questions about sales performance, customer behaviour and product preferences, with pandas for the analysis and Matplotlib for the charts.
Pandas · Matplotlib · Jupyter
One of my first projects. A text editor that runs in the terminal, with open, save, undo and redo, written in plain Python without outside libraries.
Python
GET/skills
GET/education
Patan Multiple Campus, Tribhuvan University
GPA 3.55 / 4 · First runner-up at Ascol Hackathon 2024
Kailali Model
GPA 3.05 / 4
GVN
GPA 3.85 / 4
GET/contact
I'm looking for backend and data engineering work, and I still take freelance projects in scraping, automation, bots and APIs. Email is the quickest way to reach me. I'm on Nepal time, UTC+5:45.
