HARSHAL PALANDE HARSHAL PALANDE HARSHAL PALANDE
Hello
Harshal is a passionate developer and AI enthusiast. I love building smart, real-world tech solutions—from web apps to automation projects. Always curious, always learning.
Projects⤵
Pack My Bag
Smart PackingAdds Gear by Weather
Live WeatherReal-Time Updates
Fast.Clean.Custom List
More Details →Curious huh!
- Problems I Faced
- 1. Dynamic Weather Integration: Fetching real-time weather data using OpenWeather API and handling async + errors.
- 2. Smart Packing Logic: Building condition-based packing logic using weather mappings — tricky but fun!
- 3. Flask Backend Setup: Routing, API endpoints, Railway deployment, and environment setup.
- 4. UI/UX Design: Iterating on a clean but modern layout that’s also mobile-friendly.
- 5. API Rate Limits: Learned to manage fallback behavior when API limits are hit.
Flight Delay Prediction
Real flight data usedTrained for accuracy
Track real-time flights.Anywhere. Anytime.
Know airport insights.Like a local.
More Details →Curious huh!
- Problems I Faced
- 1. Working with Limitations: The AviationStack API was powerful, but limited — I had to stay creative with what I could offer while keeping the app useful and focused.
- 2. The API didn’t always return results : Sometimes I got empty data or errors, and I wasn’t sure if it was my code or the API’s fault.
- 3. Gradio UI Constraints : Using Gradio gave me a fast way to launch the interface, but it came with some customization limits — especially around styling and layout flexibility..
- 4. Designing for Simplicity : It was tempting to add more features, but I learned that simplicity often serves users better
- 5. Integrating Two Different APIs Smoothly : Managing responses from both AviationStack and OpenSky Network — with different formats, structures, and rate limits — made coordination tricky.
House Expense Tracker
Track. Save. Repeat.Real-time insights. Smart forecasts.
Smarter spending.Manage, predict & control your expenses.
🧠 AI InsightDetect changes. Plan smarter spending
Just upload it.CSV to clarity in seconds.
More Details →Curious huh!
- Problems I Faced
- 1. Running the whole notebook over and over was exhausting: If I refreshed the page or forgot to save progress, I had to rerun everything from scratch. And when it took time to install Prophet, I started appreciating the “Save to Drive” feature a lot more.
- 2. Prophet was slow to install and run : Especially on Colab, it felt like Prophet took years to install. And even when it worked, running forecasts took a noticeable chunk of time. I learned patience and caching 😅
- 3. Uploading the CSV took longer than expected: The first time I tested with a full expense file, it just… sat there. I thought it broke. Then I realized Prophet and pandas were just processing. That taught me I should show a loader or message while the file’s being read.
- 4. Some categories didn’t have enough data to predict : I tested with my “shopping” category — only 3 rows. The model struggled. I had to add a check to warn when there’s not enough data to forecast properly.
Truthlens
TruthLens at Work.Unmask bias. Reveal truth.
Scan the News.Bias, tone & facts decoded.
Choose Your Lens.Professor, Detective, or Neutral view.
One Click Away.Paste a link. Know the truth.
More Details →Curious huh!
- Problems I Faced
- 1. Large Model Dependencies: Loading multiple heavy Transformer models caused memory issues and slow inference on local environments.li>
- 2. Article Extraction Failures: Some websites blocked newspaper3k parsing or returned incomplete content. Needed robust error handling.
- 3. Bias Detection Challenges: Zero-shot classification models can return misleading labels when context is vague or the article is too short.
- 4. Summarization Quality: The summary often required fine-tuning to avoid overly generic or incomplete sentences.
HTML
CSS
JavaScript
React
Python
Flask
MongoDB
Git
Figma
OpenAI
VS Code
HTML
CSS
JavaScript
React
🧑🏻💻 Experience
Junior Developer | Data Science Enthusiast @ Atlee
Feb 2024 – March 2025
I worked on Atlee’s dance class mobile app — built to help users discover, schedule, and join dance sessions with ease.
As a Junior Developer and Data Science Enthusiast, I contributed to the app’s stability, performance, and overall user experience. My work involved:
- Documentation: Wrote clean, easy-to-follow technical documents to help the team stay organized and speed up development.
- Testing: Ran detailed tests to catch bugs early, improve stability, and make the app feel more polished.
- Debugging models: Refined app-side models to improve accuracy and speed, especially in performance-critical areas.
- Team collaboration: Worked closely with developers and testers to ship features smoothly and fix issues efficiently.
This role gave me hands-on experience building for real users and taught me how to balance clean code with fast iteration in a collaborative team setting.
Exploring the intersection of AI, Data Science & Computer Science — one line of code at a time.
📄 Download CV