Summary
Overview
Work History
Education
Skills
Certification
PROJECTS
Timeline
Generic

HIMANSHU GURJAR

Bengaluru

Summary

Senior Machine Learning Engineer with 5+ years of experience in the end-to-end design, development, and deployment of Generative AI solutions.

Overview

5
5
years of professional experience
1
1
Certification

Work History

SENIOR MACHINE LEARNING ENGINEER

goML
01.2024 - Current
  • Contributed to AI solutions for automated document generation and user support, reducing average document preparation time by an estimated 50% and decreasing user support tickets related to document queries by approximately 10-15%.
  • Developed LLM & RAG-based chatbots, achieving an initial first-contact resolution rate of over 70% for common FAQs, handling an average of 500+ daily user interactions, and leading to a reduction in escalation to human agents by approximately 20% for covered topics.
  • Designed and implemented scalable data pipelines for information retrieval, successfully supporting an initial cohort of 100+ active users and processing several GBs of new data monthly with a target uptime of 99.5%, ensuring data isolation.
  • Developed AI agents to translate natural language into NoSQL queries, improving data retrieval efficiency and accuracy.
  • Built robust classifiers using LLMs and NLP techniques (few-shot learning) for effective query categorization and domain-specific information extraction.
  • Managed persistent chat histories, enhancing user experience and contextual understanding in conversational AI applications.
  • Technologies: Python, Open Search, RAG, DynamoDB, LLM agents, NLP, Few-shot learning, Vector Databases, Open AI, Fast API

SENIOR MACHINE LEARNING ENGINEER

BLUEVISION
01.2020 - 01.2024
  • Led development of computer vision models for agricultural commodity classification, improving quality inspection and reducing manual labor.
  • Implemented Mask R-CNN for object segmentation and homography for image data alignment.
  • Developed and fine-tuned ResNet50-based classification models with hierarchical techniques, achieving a 7% improvement in accuracy for fine-grained image categorization.
  • Improved image data quality via computer vision, boosting prediction accuracy by 10%.
  • Spearheaded data annotation and developed custom Spacy Named entity recognition models achieving 96% accuracy.
  • Built ensemble classification models (Word2Vec, TF-IDF, fine-tuned BERT) for email analysis, attaining 93% accuracy.
  • Integrated OCR for attachment text extraction into classification/NER pipelines.
  • Collaborated with cross-functional teams to deploy CV and NLP solutions to production.
  • Technologies: Python, TensorFlow, tf-serving, Docker, Flask, Computer Vision, Mask R-CNN, ResNet50, NLP, Spacy, BERT, Tesseract OCR, TF-IDF, Word2Vec

Education

Bachelor of Technology - Computer Science Engineering

ICFAI
Hyderabad
01.2018

Skills

  • Programming Languages & Core Data Libraries: Python, SQL
  • LLM & AI Development Technologies: LangChain, OpenAI, RAG (Retrieval Augmented Generation), Prompt Engineering, Computer Vision, ML System Design
  • Cloud Platforms & Services: AWS (Sagemaker, Lambda, Bedrock, OpenSearch), Azure
  • Databases, MLOps & Development Tools: Qdrant (Vector Database), MongoDB, DynamoDB, TensorFlow Serving, Fast API, Docker, Git

Certification

  • LLM101x: Large Language Models: Application through Production, Databricks, 2024-02-27
  • LLM102x: Large Language Models: Foundation Models from the Ground Up, Databricks, 2024-05-29
  • Deep learning Nano Degree, Udacity, 2019-01-01

PROJECTS

PROJECTS

Synthetic Symphony: AI-Powered Drug Discovery Platform | GoMilers (Team) - 1-Week Hackathon

  • Developed an innovative solution to accelerate early-stage drug discovery, tackling critical speed, cost, and failure rate issues in traditional methods.
  • Engineered a pipeline for protein binding site prediction and AI-driven ligand recommendation.
  • Integrated key bioinformatics tools and machine learning models:
    AlphaFold DB & Fpocket for protein structure sourcing and binding pocket identification.
    Biopython for protein/pocket structure analysis.
    ESM-2 for sequence-based protein insights.
    TxGemma (Generative AI on SageMaker) for predicting binding sites and generating ligand candidates.
  • Showcased potential to significantly reduce time-to-clinic, lower wet-lab experiment costs, and identify novel targets on previously "undruggable" proteins.

Timeline

SENIOR MACHINE LEARNING ENGINEER

goML
01.2024 - Current

SENIOR MACHINE LEARNING ENGINEER

BLUEVISION
01.2020 - 01.2024

Bachelor of Technology - Computer Science Engineering

ICFAI

Similar Profiles

CREATE PROFILE
HIMANSHU GURJAR