Home on Ujjal

Software Engineer

Tue, 01 Jan 2019 00:00:00 +0000

Openmynz (Client: ATX)
Project-Based Engagement

Worked on video distribution and conversion platforms (DVIS-2, MDmini) for enterprise clients.
Built and optimized backend services using Python, Django, WebSockets, and gRPC.
Focused on performance optimization, including memory leak detection, CPU usage reduction, and system stability improvements.
Improved TimescaleDB and PostgreSQL performance through query optimization and schema tuning.
Deployed and maintained services using Docker, Docker Compose, and Kubernetes on Linux environments.

Tech Stack: Python, Django, WebSocket, gRPC, Docker, Kubernetes, TimescaleDB, PostgreSQL, Linux

AI Engineer

Sat, 01 Jul 2023 00:00:00 +0000

Encubate Tech Private Ltd. · Pune, India
July 2023 – Present

Built a resume ranking and retrieval system using RAG architecture and fine-tuned BERT-based models to improve job–candidate matching accuracy.
Generated resume and job description embeddings using Sentence Transformers and optimized ranking with cosine similarity–based loss.
Designed a resume segmentation system using a Bi-LSTM model to classify resume sections (experience, education, skills, etc.), improving downstream parsing accuracy and inference efficiency.
Implemented NER pipelines with BERT for extracting personal and structured information from resumes, deployed as scalable APIs.
Contributed across the end-to-end ML lifecycle: data preparation, model fine-tuning, evaluation, deployment, and monitoring.
Deployed and managed AI services on AWS and Azure with production-grade reliability.

Tech Stack: Python, NLP, PyTorch, Transformers, RAG, Sentence Transformers, MongoDB, AWS, Azure

AI Application Engineer

Mon, 01 Jan 0001 00:00:00 +0000

Encubate Tech Private Ltd. · Pune, India

All About Docker and Docker compose

Sat, 05 Apr 2025 00:00:00 +0000

Into to Docker and Docker compose

Paper Implementation — Learning Text Similarity with Siamese Recurrent Networks

Fri, 01 Sep 2023 00:00:00 +0000

Fine-Tuning Sentence Transformer Models: A Case Study

Tue, 01 Aug 2023 00:00:00 +0000

Fine-Tuning Sentence Transformer Models.

RNN (Recurrent Neural Network)

Tue, 27 Jun 2023 00:00:00 +0000

Text generation with RNN

LSTM (Long Short-Term Memory)

Tue, 30 May 2023 00:00:00 +0000

LSTM (Long Short-Term Memory)

LSTM is a type of RNN architecture that is widely used for sequnce modeling task
LSTM overcome RNN limitation(vanising gradient) by introducing a memory cell and three gating mechanisms.
Memory cell in LSTM allows to store and access information over long sequence
LSTMs use a series of gates which control how the information in a sequence of data comes into, is stored in and leaves the network. they are -
- forgot gate
- input gate
- output gate

Application

NLP task- named entity recognition, sentiment analysis, machine translation etc.
Speech Recognition - automatic speech recognition, speech-to-text conversion etc.
Time Series Analysis and Forecasting - stock market prediction, weather forecasting etc.

Architecture

Forget gate layer

First step in the process is Forgot gate. This gate telling the LSTM how much information keep from previous state. Output of this gate is between 0 and 1. Output of this forgot gate multiply with previous LSTM output. output of forgot gate is 0 implies -> Forget all previous memory output of forgot gate is 1 implies -> Keep all previous memory output of forgot gate is 0.5 implies -> Keep some of previous memory

Logistic Regression

Tue, 23 May 2023 00:00:00 +0000

All about Logistic Regression

Python Advance

Wed, 25 Jan 2023 00:00:00 +0000

Post description

About Me

Mon, 01 Jan 0001 00:00:00 +0000

About me page.

Resume

Mon, 01 Jan 0001 00:00:00 +0000

Download My Resume | View on Google Drive