Background Paths
Background Paths

AJ

AkshatJha

Software Engineer & AI/ML Enthusiast

Building intelligent solutions and automating the future, one line of code at a time

About Me

Driven by curiosity and powered by code

One word which drives me - "Curiosity"

I'm an aspiring software engineer currently pursuing academics in AI and Robotics at IIT Mandi. With a strong foundation in full-stack development and applied machine learning, I thrive on building real-world applications that make a difference.

From voice bots to AI agents, I enjoy creating intelligent solutions that solve complex problems. My passion for automation drives me to build systems that not only work efficiently but also continuously improve themselves.

IIT Mandi • AI & Robotics

Innovation Focus

Building cutting-edge AI agents and automation tools that streamline complex workflows and enhance user experiences.

Passion Projects

From geospatial analysis tools to voice AI systems, I love tackling diverse challenges that push the boundaries of technology.

Continuous Learning

Always exploring new technologies and methodologies to stay at the forefront of AI/ML and software development.

Experience

Building innovative solutions and driving measurable impact

HyperQube Ionic

SDE Intern

RemoteJune 2025 – July 2025

Key Achievements

  • Architected a scalable analytics system with Amplitude and Clerk integration, engineered to capture live user events across various org segments and drive insights from 5K+ monthly actions.
  • Built a native meeting feature directly into the platform using Nylas and Twilio APIs, reducing manual scheduling time by 30% and increasing user retention by 50%.

Technologies Used

Data/Business Analytics
Amplitude Integration
Event Tracking
REST APIs
Twilio API
Nylas Calendar API

VectorX DB (prev. LaunchX Labs)

SDE Intern

Bengaluru, KarnatakaDec. 2024 – Jan. 2025

Key Achievements

  • Designed AI agents (LLM Agent & RAG Agent) with PGVectorStore and GPT-4o-mini, reducing response time by 50% and improving retrieval accuracy
  • Built a metadata extraction pipeline with LangChain's LLM, automating document processing for 1M+ documents monthly
  • Developed a dynamic filter system with FilterAgent and Claude 3.5 Sonnet, generating 100K+ JSON filters daily and boosting DB query efficiency by 40%

Technologies Used

Python
LangChain
GPT-4o-mini
Claude 3.5 Sonnet
PGVectorStore
AI Agents
RAG

Urbanchat.ai

Backend Developer Intern

Remote (Malta, Europe)June 2024 – Aug. 2024

Key Achievements

  • Deployed Kubernetes rolling update process on Amazon EKS for voicebot microservices, cutting average deployment-related downtime from 5 minutes to under 30 seconds.
  • Architected AppSumo webhook and OAuth integration, developing REST APIs to automate 95% of onboarding and reduce user setup time
  • Migrated the voicebot's speech synthesis from Amazon Polly to Deepgram, decreasing average response time from 7 seconds to 1.2 seconds, which amplified user engagement by 24%
  • Engineered the backend for an interruption feature in a voice AI chatbot using Twilio's API, cutting response delays by 35%

Technologies Used

Node.js
REST APIs
Amazon Polly
Deepgram
Twilio API
OAuth
Webhooks

Projects

Building innovative solutions across AI/ML, web development, and robotics

Light Pollution Explorer
March 2025
3rd Place - Orion Hackathon 2025

Light Pollution Explorer

Interactive web tool to visualize light pollution severity across 28 Indian states with real VIIRS satellite data

Key Achievements

  • Built interactive visualization using React.js and Node.js
  • Automated shapefile generation and per-state light intensity extraction via Python
  • Enabled 10-year trend plots with 40% reduction in preprocessing time

Tech Stack

React.js
Node.js
Python
GeoPandas
Rasterio
Web Development
Data Visualization
Dealora (IIT Mandi's Marketplace)
February 2025
3rd Place - KrackHack 2.0 (GDG)

Dealora (IIT Mandi's Marketplace)

Mobile/web platform for second-hand item exchange at IIT Mandi with real-time updates and authentication

Key Achievements

  • Built a mobile/web platform for second-hand item exchange at IIT Mandi using React (TypeScript)
  • Developed backend with Node.js and Supabase (PostgreSQL) for real-time updates and auth
  • Won 3rd place in KrackHack 2.0 (GDG)

Tech Stack

PostgreSQL
React
Node.js
Express
TypeScript
Supabase
AI-Powered Pulsar Detection
January 2025
3rd Place - SpaceCode Hackathon

AI-Powered Pulsar Detection

Enhanced pulsar detection system achieving 97.35% accuracy using advanced machine learning techniques

Key Achievements

  • Enhanced pulsar detection to 97.35% accuracy using Random Forest by optimizing features and addressing class imbalance
  • Designed a CNN-based pulsar classification pipeline in PyTorch with 4 convolutional layers
  • Achieved 3x faster training with batch normalization

Tech Stack

Python
Seaborn
Scikit-Learn
PyTorch
Jupyter Notebook
SustainX
Aug 2024 - Dec 2024
Top 10 teams - DP Open House

SustainX

Flutter-Firebase app that tracks real-time energy use, air quality, and carbon footprint, targeting college students and campuses

Key Achievements

  • Developed a Flutter-Firebase app that tracks real-time energy use, air quality, and carbon footprint, targeting college students and campuses
  • Integrated Gemini chatbot for interactive sustainability guidance, alongside custom UI, Google Fonts, and smart device control features

Tech Stack

Flutter
Dart
Firebase
Gemini API
IoT
Telecommunications Using ROS
January - May 2024

Telecommunications Using ROS

Robot teleoperation system using ROS for communication and remote command control

Key Achievements

  • Designed and implemented a robot teleoperation system using ROS for communication
  • Integrated Arduino (C++) for hardware control and Python (Tkinter) for user interface
  • Developed a mini UI to send commands via a ROS-based server with NodeMCU execution

Tech Stack

ROS
Python
Tkinter
Arduino
C++
NodeMCU
SproutVestor
February 2024

SproutVestor

Web application to assist novice investors and students in making informed stock selections

Key Achievements

  • Built a web application using Next.js, React and Appwrite (Backend as a Service) to assist novice investors
  • Incorporated an initial questionnaire to gather user's investment timeframes and preferences
  • Developed a personalized stock recommendation feature based on user responses with secure access

Tech Stack

Next.js
React
Appwrite
TypeScript
Authentication

Technical Skills

Comprehensive toolkit for building modern, scalable applications

Programming Languages

Python
JavaScript/TypeScript
C++
Java
Golang

Libraries & Frameworks

TensorFlow
PyTorch
Scikit-Learn
React.js
Next.js
Node.js
Express.js
Flask
Django

Tools & Platforms

Git
Docker
Kubernetes
AWS
GCP
Supabase
Salesforce

Data & Analytics

NumPy
Pandas
Matplotlib
Seaborn
PostgreSQL
MongoDB
GeoPandas
Rasterio

Achievements

Recognition for innovation and technical excellence

Hackathon Wins

Orion Space Hackathon 2025

3rd Place

Light Pollution Explorer project

SpaceCode Hackathon

3rd Place

AI-Powered Pulsar Detection

KrackHack 2.0 (GDG)

3rd Place

Dealora Marketplace project

Certifications

Google Cloud Public Profile

Comprehensive cloud skills and achievements

Advanced Python: Working with Data

LinkedIn Learning Certification

DevOps with AWS

LinkedIn Learning Certification

Career Essentials in Generative AI

Microsoft and LinkedIn Certification

Get In Touch

Let's collaborate on your next project or discuss opportunities

Contact Information

Location

Everywhere

Resume

Download my latest resume to learn more about my experience and skills.

Download Resume