Akshay Aralikatti

Building intelligent systems that bridge AI, robotics, and real-world impact.

Akshay Aralikatti

Projects

Voice Bot

Voice-Controlled Robot

Edge-deployed AI robot on Raspberry Pi 5 with Picovoice wake-word (<10ms), OpenAI Whisper ASR, and Gemma-3-1B intent parsing. MediaPipe + OpenCV human-following at 5 FPS via ROS 2 Jazzy.

ROS 2 Whisper MediaPipe Gemma
View Code
RAG Pipeline

Eluvio RAG Agent

RAG pipeline using llama3.2:3b with Selenium crawler scraping 10K+ JS-rendered blocks. Chunked and embedded via Sentence Transformers, indexed with FAISS for sub-1s retrieval.

LLaMA FAISS Flask Selenium
View Code
ADAS System

ADAS System

Real-time C++ OpenCV pipeline detecting lanes (RANSAC, Kalman Filter), vehicles (Haar Cascades), traffic signs & pedestrians (YOLO). Sub-40ms latency for high-speed driving.

C++ OpenCV YOLO ROS
View Code
Toy Twitter

Toy Twitter

Twitter's recommendation algorithm simulation on synthetic data. Bipartite graph construction, SALSA-style random walks, and feature-based tweet scoring with author bonuses.

Python Graph Algorithms Docker
View Code
EcoLocator

EcoLocator

Dockerized Django app on GCP with PostGIS mapping 150+ composting sites via OSM Overpass API. React+Bootstrap UI with authentication and impact tracking leaderboard.

Django PostGIS React GCP
View Code
Incremental Cache

Incremental Retrieval Cache

Production-ready AI system with pluggable storage (SQLite/MongoDB Atlas). Features drift decisioning, novelty filtering, and budget-aware assembly for optimal retrieval.

Python FastAPI MongoDB Docker
View Code
Vibe Tracker

Vibe Tracker

Desktop health app built at CalHacks 11.0. Real-time slouch detection and emotional expression analysis using MediaPipe and Hume AI for personal wellness monitoring.

Electron MediaPipe Hume AI
View Code

Experience

Research Assistant – ML, Bioacoustics

Elephant Listening Project · California State University, Chico
Aug 2025 – Present
  • Owned end-to-end inference validation pipeline for elephant-audio models: Keras → ONNX → TFLite (CPU/XNNPACK) → TensorRT FP16/INT8 on NVIDIA Jetson Orin; designed conversion ladder to isolate accuracy vs. quantization regressions across runtimes and hardware targets.
  • Designed runtime qualification workflows comparing FP32/FP16/INT8 artifacts against holdout datasets; benchmarked ~21K clips for accuracy, precision/recall, AUC, avg/p95 latency, and peak RAM across CPU (ONNX Runtime) and GPU (TensorRT).
  • Implemented INT8 TensorRT calibration (IInt8EntropyCalibrator2) with holdout sampling; profiled quantization-induced metric shifts at fixed operating thresholds to harden inference reliability across custom accelerator targets.
  • Hardened deployment preprocessing (training-aligned spectrogram/STFT pipeline) with a stable I/O contract across all runtimes; packaged inference engines, calibration caches, metrics tables, and a reproducible Jetson runbook.
  • Built scalable preprocessing pipelines for 4,750+ hours (300GB+) of audio data; ran distributed model training and inference benchmarking on SDSC Expanse HPC (multi-GPU).

Software Developer

Virtual Labs – IIIT Hyderabad
Nov 2022 – Mar 2024
  • Led design of 7 interactive OS labs using MERN stack, serving 8.5M+ participants across 46M+ page views
  • Refactored 800+ lines legacy code, integrated Jest/Cypress tests, improving stability by 30%
  • Built REST APIs with MongoDB/PostgreSQL persistence and GitLab CI/CD pipelines

Full-Stack Engineer

Indiahikes
Sept 2017 – Dec 2022
  • Built core booking platform handling 20,000+ trekkers and 35,000+ bookings annually
  • Containerized services using Docker, deployed cloud-native microservices on AWS/GCP
  • Implemented caching via Redis, reducing booking-flow latency by 40%
  • Led 80+ high-altitude Himalayan expeditions while gathering user feedback for product improvements

UI/UX Design Intern

Wissen Technologies
Apr 2024 – Jul 2024
  • Designed 6 production-ready web layouts in Figma, improving navigation efficiency by 30%
  • Collaborated with engineers to translate wireframes into responsive React components

Skills

Languages & Frameworks

C/C++ Python Go JavaScript Django FastAPI Node.js React Spring Boot

ML & AI

PyTorch TensorFlow LLM Fine-tuning RAG Pipelines TensorRT FAISS Vector DBs

Systems & Robotics

ROS 2 OpenCV MediaPipe Multithreading Low-latency Pipelines TCP/IP

DevOps & Cloud

Docker Kubernetes AWS GCP GitHub Actions Terraform Prometheus

Blogs

Coming soon — technical deep-dives and lessons from the field.