AI DEVELOPER · AMULYA

AI/ML ENGINEER · GENERATIVE AI · NLP · COMPUTER VISION

Anuj K
Janardhanan

I build AI applications across machine learning, deep learning, generative AI, NLP and computer vision. Currently developing AMDEX for recruitment operations at Amulya Corporate Services LLP, after working on satellite image super resolution at ISRO and predictive analytics at Adani.

From models to applications, I build AI systems end to end.

I'm an AI/ML Engineer and B.Tech CSE graduate from Indus University, working across machine learning, deep learning, computer vision, NLP and generative AI. As an AI Developer at Amulya Corporate Services LLP, I'm developing AMDEX, a recruitment operations and candidate search platform that centralizes resumes and structured candidate records, combining database workflows with AI-assisted retrieval and matching. Previously, I built ESRGAN and SRDiff models in PyTorch at ISRO's Space Applications Centre and worked on churn prediction at Adani Enterprises. My work spans model development, data/ML pipelines, backend AI systems and deployment tooling.

CGPA9.16 / 10
GRADUATEDJUN 2026
CURRENT ROLEAI Developer
Amulya Corporate Services LLP
FOCUSAI/ML EngineeringDeep Learning & Computer VisionGenerative AI & NLP

What I build with.

AI / ML FRAMEWORKS

PyTorchTensorFlowscikit-learn XGBoostHugging Face TransformersFastEmbed

DEEP LEARNING & CV

CNNsGANsDiffusion Models ESRGANSRDiffResidual Networks SISRAttentionPSNR / SSIM

GENERATIVE AI / NLP & SEARCH

LLMsOllamaSemantic Search Information RetrievalVector SearchEmbeddings RerankingPrompt EngineeringText ClassificationEmotion Detection

BACKEND & DATABASES

FastAPINode.jsFlask PostgreSQLpgvectorMongoDB AtlasMySQL

PROGRAMMING & WEB

PythonC++JavaJavaScript TypeScriptReactNext.jsTailwind CSS HTMLCSSJSPServlets

INFRASTRUCTURE & TOOLS

DockerRedisCelerynginx PrometheusGrafanaGit GitHubJupyterVS Code

Where I've put this to work.

JUL 2026 — PRESENT

AI Developer

Amulya Corporate Services LLP

  • Engineering AMDEX, a recruitment operations and candidate search platform that centralizes resumes and structured candidate profiles. Currently in development/testing, with a design scale of 5,000+ resumes and role-based access for recruiter workflows.
  • Built candidate search and retrieval workflows using candidate details, skills, experience and PostgreSQL full text search, enhanced by pgvector semantic similarity, JD-based matching and cross-encoder reranking.
  • Developing automated resume processing for PDF/DOCX extraction, OCR and embeddings using FastEmbed, with asynchronous processing through FastAPI, Redis and Celery.
  • Building the platform with a Next.js frontend, FastAPI backend and PostgreSQL with pgvector. Containerized with Docker, with nginx, Prometheus and Grafana supporting deployment and monitoring during testing.
DEC 2025 — JUN 2026

Project Trainee - Deep Learning & Computer Vision

Space Applications Centre, ISRO

  • Built ESRGAN and SRDiff super resolution models from scratch in PyTorch to sharpen low resolution satellite imagery, tuning perceptual loss functions and validating output with PSNR and SSIM.
  • Worked on SRDiff which is a diffusion based architecture using stochastic denoising to recover fine detail that standard upscaling methods miss.
  • Owned the full training pipeline independently, from data preprocessing and CNN architecture with attention and multi scale feature extraction to training loops and iterative debugging.
JUL 2025 — SEP 2025

Analytics & AI Intern

Adani Enterprises Limited

  • Built a churn prediction model for the ICM project using XGBoost and K-Means clustering to flag at risk customers, delivering AUC-ROC evaluated churn scores to the analytics team.
  • Ran the full ML workflow in Python, from cleaning raw data and engineering features to tuning hyperparameters, and translated model outputs into plain language retention recommendations.

From language models to full stack applications.

WorkSync AI

INTELLIGENT HR MANAGEMENT SYSTEM

A chat assistant built on a local LLM (Ollama) that lets HR staff query employee data in plain English no SQL needed wired into a React/TypeScript frontend. RESTful APIs in Node.js, backed by MongoDB Atlas, handle real time data operations, access control and full stack integration.

ReactTypeScriptNode.js MongoDB AtlasTailwind CSSOllama
View repository →

Moodify

AI EMOTION DETECTION WELLNESS APP

An NLP classifier detects user emotion from free text input and maps it to a mood persona, then hands off to a local generative model for prompt engineered follow up conversation. Full Flask app with auth, sessions and a UI that shifts dynamically with detected mood built modular so a fine tuned transformer can slot in later.

PythonFlaskscikit-learn NLPOllamaGenerative AI
View repository →

Weekendy

HOTEL BOOKING & ITINERARY PLATFORM

A hotel booking platform with separate user and admin roles, a clean responsive interface and end to end CRUD booking workflows built with JDBC and prepared statements including date range validation and real time room availability checks.

JSPServletsHTMLCSSApache Derby (JDBC)
View repository →

Education & certifications.

Indus University, Ahmedabad

B.Tech, Computer Science Engineering : 2022 - 2026

CGPA 9.16 / 10 · Relevant Coursework: Machine Learning, OOPS, Data Structures & Algorithms, Database Systems

Shree Narayana Guru Vidhyalaya

Secondary & Higher Secondary Education : 2010 - 2022

CERTIFICATIONS & ACHIEVEMENTS

  • AWS Academy Machine Learning Foundations AWS ACADEMY · AUG 2025 · 20 HRS View Credly badge →
  • Data Analysis Using Python IBM · JUN 2026 View Credly badge →
  • ISRO Project Trainee at Space Applications Centre COMPETITIVELY SELECTED National-level deep learning research program on satellite image super-resolution.

Let's build useful AI systems.

Working on machine learning, deep learning, generative AI, NLP or computer vision? I'd love to connect and talk about the problems you're solving.

AHMEDABAD, GUJARAT, INDIA