Skip to content
View lakshiaharan's full-sized avatar

Block or report lakshiaharan

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.

Content in all repositories owned by your account will be closed.
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
lakshiaharan/README.md

Hi, I'm Lakshiaharan πŸ‘‹

Undergraduate engineer focused on autonomous systems simulation, audio deep learning & digital signal processing, and cloud-native backend architectures.

GitHub LinkedIn Live Demos


πŸš€ Featured Projects

Interactive autonomous robotics simulation & pathfinding benchmark suite
React 19 TypeScript Node.js WebSockets Tailwind CSS Vitest

  • Simulation Engine: 50 Hz telemetry stream over WebSockets with analytical 2D LiDAR raycasting & differential-drive unicycle kinematics.
  • Pathfinding Benchmarks: $A^*$ (0.62–0.94 ms), Dijkstra (2.01–3.77 ms), Greedy BFS (0.12–0.27 ms), and continuous RRT trajectory sampling.
  • Verification: 26/26 automated unit and physics integration tests passing with Vitest.
  • πŸ”— Live Demo β€’ Source Code

Deep learning & signal processing framework for environmental sound classification
Python PyTorch Librosa Spectrogram Analysis Coordinate Attention

  • Signal Engineering: Multi-resolution STFTs combined with differential spectral velocities ($\Delta, \Delta^2$) to capture acoustic transients.
  • Model Architecture: Coordinate Attention CNN backbone evaluated on the ESC-50 dataset (2,000 clips across 50 categories).
  • Evaluation: 5-fold cross-validation with a strict leak-free split methodology (76.80% Β± 1.45% CV accuracy).
  • πŸ”— Web Demo β€’ Source Code

Decoupled multi-tier cloud prototype built on AWS services
AWS (EC2, DynamoDB, SNS, S3, CloudWatch) TypeScript Express PM2

  • Cloud Architecture: REST API on EC2 Ubuntu Linux, DynamoDB On-Demand tables, Amazon SNS asynchronous pub/sub event fanout.
  • Observability & Security: Custom CloudWatch metric dispatcher with CPU alarms; IAM Instance Profiles with zero hardcoded credentials.
  • πŸ”— Live Web App β€’ S3 Static Site β€’ Source Code

πŸš‘ SynapseAid

Browser-based emergency response simulation with deterministic multi-agent matching & hybrid retrieval
TypeScript React Node.js Spatial GIS Gale-Shapley Matcher

  • System Design: Dual-stream conversational triage integrating speech synthesis, retrieval, and GIS vector distance matching.
  • Custom Algorithms: In-browser Gale-Shapley responder matching, 24-D vector projection + BM25 rank fusion, and SHA-256 cryptographic audit ledger.
  • πŸ”— Live Demo β€’ Source Code

πŸ› οΈ Technical Skills

  • Languages: TypeScript, JavaScript, Python, C/C++, SQL, HTML5/CSS3
  • Frameworks & Libraries: React 19, Node.js, Express, PyTorch, Librosa, Vitest, Tailwind CSS
  • Cloud & DevOps: AWS (EC2, DynamoDB, SNS, CloudWatch, S3), Vercel, Docker, Git, Linux
  • Core Engineering: Autonomous Kinematics, Path Planning ($A^*$, RRT), Digital Signal Processing (STFT), Event-Driven Architecture

Popular repositories Loading

  1. aegis-kinetic-studio aegis-kinetic-studio Public

    High-performance interactive robotics and kinetic telemetry simulation platform with real-time pathfinding algorithms, physics engine, and telemetry HUD.

    TypeScript

  2. synapse-aid synapse-aid Public

    Full-stack AI-driven emergency triage and assistance platform with real-time vector visualizer, conversational triage orchestrator, and voice interface.

    TypeScript

  3. spice-route-restaurant spice-route-restaurant Public

    Decoupled multi-tier restaurant operations platform with TypeScript Express API on AWS EC2, DynamoDB NoSQL persistence, SNS messaging, CloudWatch metrics, and S3 static hosting.

    JavaScript

  4. audio-classification-esc50 audio-classification-esc50 Public

    End-to-end deep learning framework for environmental sound classification on ESC-50 with multi-resolution STFTs, delta spectral velocities, and coordinate attention.

    Python

  5. lakshiaharan lakshiaharan Public

  6. aero-scan aero-scan Public

    Python