UTYL
Labs
An engineering portfolio showcasing hobby projects and experiments.
Stocks
The Problem
Building a trading strategy requires complex integrations and expensive tools for testing in simulated environments.
The Solution
Developed a comprehensive trading platform that simulates real-time trade execution, allowing users to backtest and refine strategies without financial risk.
Key Highlights
- Real-Time Stock Trading API Integration
- Comprehensive Portfolio Tracking
- Detailed Trade History Retrieval
- Strategy Simulation & Backtesting
Itinerary
The Problem
Travel planning involves juggling multiple apps for flights, hotels, and recommendations, leading to an organized but chaotic experience.
The Solution
Engineered an all-in-one planner that aggregates cheapest flights, hotel options, and localized culture guides into a unified interface.
Key Highlights
- Personalized Travel Itineraries
- Flight and Hotel Aggregation
- AI-Powered Activity Recommendations
- Localized Culture and Language Support
Paperpal
The Problem
Reading lengthy research papers is time-consuming and inaccessible for on-the-go learning or visually impaired individuals.
The Solution
Built an AI pipeline that extracts text from PDFs, generates concise summaries, and synthesizes them into speech running entirely on free-tier services.
Key Highlights
- Seamless PDF Text Extraction
- Text-to-Speech Synthesis
- AI-Generated Content Summarization
- Natural Language Q&A
Uptime
The Problem
Existing uptime monitors are expensive or bloated, lacking a simple pay-as-you-ping model for small projects.
The Solution
Created a scalable, lightweight monitoring service with a distributed worker architecture that checks site health and dispatches instant alerts.
Key Highlights
- Distributed Health Checks
- Realtime Email & Slack Notifications
- Customizable Monitoring Intervals
- Detailed Analytics and Reporting
Media Search Engine
The Problem
LLMs have limited context windows and only process text, making it difficult to natively perform question-answering over large video documents.
The Solution
Built a highly-scalable event-driven microservices architecture that automatically extracts audio, transcribes speech, and creates embeddings for semantic QA.
Key Highlights
- Seamless Audio Extraction & Transcription
- Semantic Search via Text Embeddings
- Event-Driven Communication with Apache Kafka
- Real-time Job Polling with Redis Cache
Concurrent Download Manager
The Problem
Downloading large files sequentially over a single HTTP connection is often bottlenecked by bandwidth limits, network latency, or inconsistent network conditions.
The Solution
Leverages HTTP Range Requests to intelligently split files into multiple chunks, downloading them concurrently using goroutines while natively falling back to single-threading if required.
Key Highlights
- Blazing Fast Concurrent Downloads
- Smart Fallback Mechanism
- Real-Time Progress Tracking
- Memory Efficient syncing
Tempo
The Problem
Manually cutting out pauses from spoken videos takes hours of tedious scrubbing relative to video length, directly hurting editor and creator productivity.
The Solution
Tempo processes raw video in two passes: analyzing uncompressed audio RMS energy to precisely map silent gaps, then perfectly reconstructing the video using FFmpeg without synchronization drift or pre-roll glitches.
Key Highlights
- Automated Silence Detection via RMS
- Zero Audio/Video Sync Drift
- Glitch-free Frame Cutting
- Customizable Presets (Aggressive/Balanced/Gentle)
Vestri
The Problem
Scaling web applications horizontally requires efficient traffic distribution to prevent individual servers from becoming overwhelmed while others idle.
The Solution
Serves as a reverse proxy deploying Round Robin and Weighted Round Robin strategies, paired with active health checking to dynamically route requests and maximize capacity.
Key Highlights
- Path-based Routing
- Weighted Round Robin Load Balancing
- Active Backend Health Checking
- Declarative YAML Configuration
Building Something New?
We love tackling challenging engineering problems and building robust, scalable solutions. Let's talk about your next big idea.