Hi, my name is

Nitin Monga.

I build ML & GenAI Projects

A passionate AI developer

About Me

I am a highly skilled and results-driven Data Scientist and Generative AI Engineer with a profound dedication to leveraging advanced AI and Machine Learning technologies to tackle intricate challenges. My expertise traverses multiple domains, including Statistical Data Analysis, Machine Learning, Deep Learning, Natural Language Processing, Computer Vision, Genrative AI, Large Language Models, LangChain and Vector Databases.

Equipped with a deep understanding of these technologies, I adeptly design and implement innovative solutions that drive impactful outcomes. Proficient in programming languages such as Python, R, and SQL, I possess a robust foundation in handling vast datasets and extracting actionable insights. My proficiency extends to data visualization and communication, enabling me to effectively convey technical findings to both technical and non-technical stakeholders.

I take pride in my ability to distill complex concepts into digestible insights, facilitating informed decision-making across diverse teams and departments. Thriving in dynamic environments, I am committed to remaining at the forefront of AI technologies, continually expanding my skill set and exploring emerging trends. With a relentless pursuit of excellence, I am driven to make a meaningful impact through the application of cutting-edge AI solutions.

I am extensively engaged in delivering comprehensive training sessions tailored for prominent global IT enterprises, focusing specifically on the dynamic fields of Data Science, Machine Learning and GenAI. These sessions encompass a wide range of topics, including but not limited to Data Analysis, Predictive Modeling, Deep Learning, NLP, Vector Database, LangChain, LLMs and Development & Deployment of AI applications.

Over time, I have conceptualized and executed algorithmic trading strategies utilizing Python and AI frameworks for various trading clients in the Indian Stock Market. Utilize machine learning techniques to analyze market data and identify profitable trading opportunities. Additionally, I have adeptly provided training to non-technical individuals in effectively utilizing these algorithmic trading methodologies.

I bring forth over 19 years of diverse and extensive experience spanning the breadth of the IT domain, coupled with a dedicated focus of over 8 years specifically within the realm of Data Science and Machine Learning technology stack. Throughout my career journey, I have honed a robust skill set encompassing various facets of IT, including software development, project management, and strategic planning, while concurrently immersing myself in the cutting-edge innovations of Data Science and Machine Learning.

Data Science & Machine Learning SKills

  • Python
  • R
  • SQL
  • MySQL
  • TensorFlow 2.0
  • PowerBI
  • Tableau
  • Statistics
  • Streamlit
  • PL-SQL
  • HTML & CSS
  • Oracle Database
  • Docker
  • MLFlow

Machine Learning Algorithms

  • Data Analysis
  • Regression
  • Classification
  • Clustering
  • Forecasting
  • Computer Vision
  • Visualisations
  • NLP - Natural Language Processing
  • ANN - Artificial Neural Network
  • CNN - Convolutional Neural Network
  • RNN - Recurrent Neural Network

GenAI - Generative Artificial Intelligence

  • HuggingFace
  • LLM - Large Language Models
  • OpenAI - GPT Models
  • Embedding Models
  • LangChain
  • Vector Database
  • Pinecone
  • Chroma
  • Weaviate
  • Faiss
  • Qdrant

Financial Analysis & Equity Market Skills

  • Algorithmic Trading
  • Fundamental Analysis
  • Technical Analysis
  • Financial Modeling
  • Market Data Analysis
  • Algo Trading Strategies
  • Portfolio Optimization
  • Backtesting
  • Kite API
  • Angelone API

Experience

Technical Lead/Manager - Wipro Technologies
Oct 2011 - Feb 2014

ANZ Bank - Migrating Bank’s Servers & Software to Latest Version

  1. Managing a team of 15+ members.
  2. Determine overall project plan, structure, schedule, and staffing requirements.
  3. Actively participated in a weekly status meeting with the client, discussing project status and bottlenecks.
  4. Tracking work progress, conducting daily stand-up calls to get the daily status, andunderstanding bottlenecks or risks if any.
  5. Coordinating with offshore team, ensuring on-time deliverables.
  6. Delegating work to the team members based on the skills and keeping management informed on deliverables and raising red flags whenever saw any risk.
  7. Involved in other management activities such as – Resource Management: resource planning and loading, project planning, working on cost estimation sheet, work allocation,reporting internal and external project performance on a periodic basis, and other activities using Wipro-specific tools and matrix.

CITI Bank - Operational Risk Management

  1. Leading a team of 8 + members.
  2. Involved in requirement gathering, analysis, and feasibility.
  3. Involved in data analysis, database design, and developing and tuning critical PL/SQL objects.
  4. Technical discussion with the clients on technical design and taking inputs.
  5. Effective Monitoring, requirement, and work progress tracking.
  6. Involved in code review, design, test scenarios & Unit test cases.
  7. Working with the various streams, Project documentation, Integrated Change Control;communication with the stakeholders, and handling escalations.
Tech Lead - ACS - A Xerox Company
Jan 2009 - Oct 2011

Michlen & Office Depot DIGITAL FUEL - Implementation Digital Fuel BI tool

  1. Build & trained the technical & functional team of 10+ members.
  2. Implemented Digital Fuel Service Flow BI Reporting tool for different clients.
  3. Helping and guiding team members on the technical front whenever required.
  4. Daily status call with the client and stakeholders on status and bottlenecks.
  5. Determine overall project plan, structure, schedule, and staffing requirements.
  6. Ensuring target realizations and bulge ratios, Implementing service improvement plans,giving leads to senior management.
  7. Performed capacity planning and was responsible for project prioritization by providing information about the resource availability to business managers, project sponsors & clients.
Software Engineer - 3i infotech - India and US
Jan 2005 - Dec 2008

PSCM, John Snow Inc - Oracle BI Reports, Database

  1. Involved in requirement gathering & delivery for different Discoverer BI reports.
  2. Coordinating with the offshore team for delegation, requirement clarifications, and on-time delivery of project modules.
  3. Testing deliverables to make sure development is bugs-free and aligned with the clients requirements.
  4. Involved in Database design, developing PL/SQL objects, and creating different Oracle BI reports.
  5. Created new database PL/SQL objects such as packages, procedures, and triggers and also customized existing objects.
  6. Responsible for testing the development patches before applying to the UAT or production environment and also responsible for on-time teams deliverables.
Entrepreneurial Journey - Xilytica
Mar 2014 - Till Date

freelancing, Machine Learning, Deep Learning Generative AI, Data Science Training

  1. I have single-handedly built my training and consulting platforms, Skill Venue and Xilytica,from the ground up. Idea was to built an online learning & mentoring platform to bringtrainers, mentors & college students/ working professionals to it. So that trainers, mentorscan provide their respective services to colleges & individuals. But due to unforeseencircumstances, had to pivot the platform to providing Data Science & Machine LearningTraining & Consulting platform.
  2. While delivering tangible outcomes for my clients, I was also involved in actively developing business leads and onboarding trainers & mentors for my clients.
  3. I have also dipped my toes into digital marketing, content strategy, and running digital advertising, generating leads over various social platforms for my companies.
  4. Lastly, recruiting, and grooming talent has been a pleasant outcome of my entrepreneurial venture as I have got to be part of several peoples career journeys.
  5. Since last few years, I have trained & mentored many college students & working professionals in Data Science & Machine learning.
  6. I have been undertaking Data Science and AI projects from companies and startups in India and US.

Consulting Projects

# Vernacular Voice and Text-Based AI Agent
# Vernacular Voice and Text-Based AI Agent
Generative AI, OpenAI LLM API, Langchain, Pinecone Vector Database, Scrapping

We have developed a conversational RAG (Retrieval-Augmented Generation) application designed to assist farmers by providing answers to their queries through both voice and text interfaces. This AI assistant addresses a broad spectrum of agricultural topics, including government policies, crop cultivation practices, irrigation techniques, pest control measures, and the use of agricultural products and tools, as well as other farm-related subjects.

Proof of Concept Completion: The initial proof of concept has been successfully completed, utilizing cutting-edge technologies: OpenAI’s API for the Large Language Model (LLM), Pinecone as the vector database, and Langchain as the orchestration framework. This phase demonstrated the application’s capability to deliver accurate and contextually relevant responses to farmers' inquiries, showcasing the potential of AI in transforming agricultural support.

Next Steps: Building on this success, we are now advancing to the next stage of development. Our focus will be on training an open-source large language foundation model, which will then be fine-tuned under supervision using a dataset comprising real-world queries from farmers. This fine-tuning process will ensure that the AI model is highly specialized in addressing the unique challenges and needs within the agricultural sector.

Future Vision: In the later stages of the project, we plan to integrate a GenAI agent capable of analyzing satellite imagery to identify specific plant diseases, nutrient deficiencies, and environmental stress factors. This innovation will empower farmers to apply targeted treatments precisely where they are needed, thereby reducing chemical usage, minimizing waste, enhancing crop yield and quality, and promoting sustainable farming practices.

Complete Idea & Development done by me The entire concept was conceived and developed by me, with a focus on creating value and addressing key challenges within the agricultural sector. After thorough discussions, both the investors and the founder found this idea exceptionally compelling, recognizing its potential to significantly benefit the company. The idea not only aligns with the company’s goals but also positions us strongly in the market, driving future growth and sustainability.

Grain Quality Digitization Using CNN ResNet Model
Grain Quality Digitization Using CNN ResNet Model
Deep Learning, Convolutional Neural Network, ResNetTensorflow

The Grain Quality Digitization work package aims to create an automated system for classifying maize kernel quality using Convolutional Neural Networks (CNN), particularly the pre-trained ResNet (Residual Networks) model. This work package seeks to digitize and standardize the quality assessment process, ensuring consistent, accurate, and efficient maize classification.

Data Collection: Image Acquisition: Initial model training began with publicly available maize images. Later, we requested farmers to provide high-resolution images of maize kernels, labeled with quality grades such as A, A+ etc. These images were captured under standardized conditions to ensure uniform lighting and background. Dataset Preparation: The images were annotated based on predefined quality metrics, including kernel size, color, shape, and defects. To improve the model’s robustness and diversify the dataset, data augmentation techniques like rotation, scaling, and flipping were applied.

Model Training: The annotated maize kernel dataset was used to train the ResNet model. Using transfer learning technique, we fine-tuned a pre-trained ResNet model on our specific maize kernel dataset. Initially, the model achieved around 50% accuracy due to the limited dataset. However, as more images were added to the dataset, the model’s accuracy started to improve. Hyperparameters such as learning rate, batch size, and the number of epochs were carefully optimized to enhance the model’s performance, leading to more accurate and reliable maize quality assessments.

Inventory Analysis for a Retail Chain
Inventory Analysis for a Retail Chain
Data Analysis, Python, PowerBI, SQL Database
For a retail store client SaveMore, analysed & built sales & inventory dashboards. Sales dashboard helped supermarket clients in understanding identifying sales trends, seasonal fluctuations, top performing products, & customer buying pattern across all stores. And Inventory dashboards, helped in tracking inventory levels, demand forecasts, and reorder points. This prevents overstocking or under stocking thereby reducing inventory cost.
Brand Listening & Sentiment Analysis for Spring-Mattress Company
Brand Listening & Sentiment Analysis for Spring-Mattress Company
Machine Learning, NLP, Data Analysis, Python, PowerBI, MySQL Database
Brand Listening & Sentiment Analysis project for a client Spring-Mattress company that is based out of US. The analysis provided valuable insights into how customers perceive the brand, its products, and services. By monitoring online conversations, social media interactions, and customer reviews, the company gained a deep understanding of customer sentiment, preferences, and pain points. This information helped in identifying areas of improvement, addressing customer concerns proactively, and enhancing overall brand reputation and customer satisfaction. The actionable insights derived from the analysis guided strategic decision-making, allowing the company to optimize marketing strategies, product development, and customer engagement initiatives.
Clustered and Analysed data from HomeOwner Association (HOA)
Clustered and Analysed data from HomeOwner Association (HOA)
Machine Learning, Clustering Algorithm, Data Analysis, Python, MySQL Database
Extracted and Analysed data from HomeOwner Association (HOA) PDF documents & categorized & store the data in a SQL database, paving the way for a front-end service that provides users with insights based on the uploaded HOA documents.
Build Sales Pipeline for Cilans systems
Build Sales Pipeline for Cilans systems
Machine Learning, NLP, Scrapping, Data Analysis, Python, PowerBI, MySQL Database
For the client Cilans system who are into corporate trainings & consulting services. The project entails scraping publically available data from specified websites to collect comprehensive information about companies & working professionals, storing this data securely in a database, and utilizing Natural Language Processing (NLP) techniques for insightful analysis. The goal was to extract meaningful insights such about company, number of employees, name decision makers etc. that can benefit the client’s sales team and enhance their sales strategies. The project significantly aided the client in creating an effective sales pipeline by leveraging data scraping which led to 30% increase in leads generation. The NLP insights enabled targeted outreach and personalized communication, ensuring engagement with high-potential leads that resulted in 15-20% improvement in the conversion rate. This approach streamlined the sales process, allowing the client to prioritize opportunities and optimized follow-up actions, reducing the average sales cycle duration by 15%, resulting in faster deal closures.
Developed Machine Learning based Stocks Trading Strategies
Developed Machine Learning based Stocks Trading Strategies
Machine Learning, Time Series, Trading API, Scrapping, Data Analysis, Python, Trading Kite and AngelOne API
For a SAAS based trading platform created ML based algorithmic trading strategies. The project involved designing, implementing, and optimizing Machine Learning based algorithmic trading strategies tailored to meet the specific requirements of clients. The goal was to develop automated trading systems that can execute trades based on predefined criteria and parameters. Conducted back testing to evaluate strategy performance using historical data & optimize strategies for maximum efficiency, profitability, and risk mitigation. Backtesting strategies resulted in a quantifiable improvement in traders' profitability, with an average increase of 10% in profits across tested strategies.
Developed Smart Stocks Screeners
Developed Smart Stocks Screeners
Machine Learning, Trading Kite and AngelOne API, Scrapping, Data Analysis, Python
Developed custom trading screeners for individual retail traders & implemented these screeners on cloud servers for easier access to the traders. These screeners were designed to assist clients in identifying potential trading opportunities based on their specified filtering criteria and market conditions. The developed screeners helped traders in filtering thousands of stocks & uncover hidden gems – stocks but haven’t yet garnered wider attention. It frees up valuable time for the clients so that they can focus on in-depth analysis and trade execution. It also provides real time alerts & notifications about their trade setups price movements, or market events, allowing them to act promptly and capitalize on opportunities. Screener helped client in identifying 10 times more trading opportunities compared to when client was manually filtering & following stocks his stocks.

Personal Projects

Youtube Videos Summary and Analysis Application
NLP Huggingface OpenAI API Langchain
Youtube Videos Summary and Analysis Application
Using GenAI framework built an application that summarises & analysis YouTube Videos. The Application provides a tool for users to quickly understand and evaluate video content. It summarizes key points and highlights from videos, making it easier to grasp main ideas without watching the entire content. Used NLP features like keyword extraction, sentiment analysis, and context breakdown to offer deeper insights. Ideal for researchers, students, and busy professionals, this application saves time and enhances comprehension by distilling lengthy videos into concise, informative summaries and analyses.
Movie Recommendation Application
Collabrative Filtering Content-Based Filtering NLP FastAPI Streamlit
Movie Recommendation Application
A Movie Recommendation Application utilizes machine learning algorithms and data analytics to deliver personalized film suggestions. It processes user data, such as viewing history, ratings, and genre preferences, to generate recommendations using techniques like collaborative filtering, content-based filtering, and hybrid models. Collaborative filtering analyzes patterns in user behavior to suggest movies based on similar users' choices, while content-based filtering recommends films with similar attributes to those a user has liked.
Stock Price Prediction Using Deep Learning
Deep Learning LSTM NLP Trading APIs
Stock Price Prediction Using Deep Learning
Stock price prediction using deep learning leverages neural networks to forecast future stock prices based on historical data and market indicators. I used advanced models like Long Short-Term Memory (LSTM) networks and Gated Recurrent Units (GRUs)to capture temporal dependencies and trends in time-series data. The process involves preprocessing historical stock prices, incorporating additional features such as trading volume and sentiment analysis, and training the model on this data. The goal is to predict future price movements with improved accuracy, aiding investors in making informed decisions.
Youtube Videos Translation App
GenAI LLM OpenAI API Langchain
Youtube Videos Translation App
A YouTube video translation application for English to Hindi uses OpenAI’s language models and Generative AI (GenAI) to translate video content. It starts by extracting the English audio from the video and converting it into text using speech recognition. OpenAI’s API then translates this text into Hindi, while GenAI ensures the translation is contextually accurate and natural-sounding. The translated text is used to create Hindi subtitles or voiceovers. This process involves advanced machine learning techniques to understand and generate human-like language, making it easier for Hindi-speaking users to comprehend and engage with English-language videos.
Insurance Claims Clustering App
Machine Learning K-Means Clustering Streamlit
Insurance Claims Clustering App
An Insurance Claim Clustering App uses machine learning algorithms to categorize and group insurance claims based on various features such as claim type, severity, and fraudulent indicators. By analyzing historical claims data, the app identifies patterns and similarities, enabling insurers to streamline the claim management process. Clustering techniques, such as K-means or hierarchical clustering, are employed to segment claims into distinct categories for better handling and analysis. This helps in improving operational efficiency, detecting anomalies, and prioritizing claims that require immediate attention. Overall, the app aids in enhancing decision-making and resource allocation within the insurance industry.
200 SMA and 50 EMA Crossover Screener
Kite API Angel One API FastAPI Streamlit
200 SMA and 50 EMA Crossover Screener
The 200 SMA and 50 EMA Crossover Screener is a trading tool that identifies potential buy or sell signals based on moving average crossovers. It monitors the 200-day Simple Moving Average (SMA) and the 50-day Exponential Moving Average (EMA) of a stock or asset. When the 50 EMA crosses above the 200 SMA, it signals a potential buying opportunity, indicating an uptrend. Conversely, when the 50 EMA crosses below the 200 SMA, it suggests a potential selling opportunity, indicating a downtrend. This screener helps traders make informed decisions by highlighting significant trend shifts in the market.

Get in Touch

for any

Machine Learning, GenAI Consulting Services

or

Training Requirements

at

+91 702 294 5888

nitin@nitinai.xyz