Kanishk Kumar
Data Science Consultant | Machine Learning Engineer | Python Developer
Code smart. Think consultant.
📧 kanishkk202@gmail.com | 📱 +91-9990465116 / +91-7042693420(WhatsApp) | 📍 Noida, Uttar Pradesh
🌐 Portfolio Website | 💻 GitHub Profile | 👔 LinkedIn
🙋 About Me
Results-driven Data Science Consultant with an M.Sc. in Data Science and 2+ years of hands-on experience delivering end-to-end machine learning, NLP, and analytics solutions. Proven expertise in building automated ETL pipelines, engineering deep learning architectures (LSTMs, Transformers), and deploying scalable models using Docker and Flask. Passionate about leveraging MLOps and Generative AI to surface actionable business insights and optimize operational efficiency.
💡 Technical Skills
- Languages & Libraries: Python (Advanced), SQL (Intermediate), Pandas, NumPy, Scikit-learn, PyTorch, TensorFlow/Keras, XGBoost
- Machine Learning & AI: Natural Language Processing (BERT, GPT, Transformers), Time-Series Analysis (LSTM), Computer Vision, Statistical Modeling, A/B Testing
- MLOps & Deployment: Docker, Flask, FastAPI, Celery, MLflow, Git
- Big Data & Cloud: Apache Spark (RDD, DataFrames), Hadoop, Azure
💡 IT Skills
- Languages & Libraries: Python (Advanced), SQL (Intermediate), Pandas, NumPy, Scikit-learn, PyTorch, TensorFlow/Keras, XGBoost.
- Machine Learning & AI: Natural Language Processing (BERT, GPT, Transformers), Time-Series Analysis (LSTM), Computer Vision, Statistical
Modeling, A/B Testing.
- MLOps & Deployment: Docker, Flask, FastAPI, Celery, MLflow, Git.
- Big Data & Cloud: Apache Spark (RDD, DataFrames), Hadoop, Azure
🧠 TECHNICAL & ANALYTICAL EXPERIENCE
Independent Data Science Consultant | Self-Employed July 2025 – Present
- Real-Time Stock Price Prediction Architecture | Python, LSTM, Data Visualization
- Engineered an end-to-end time-series pipeline using LSTM networks to forecast financial asset prices at 5-minute intervals.
- Achieved an RMSE of $2.45, ensuring 88% of predictive outputs remained within 3% of the actual closing price.
- Optimized automated model retraining latency to 8 seconds, establishing production-ready low-latency infrastructure.
- 🔗 GitHub Repository
- Automated Sales Performance & Customer Analytics Dashboard | Python, SQL, Power Query
- Partnered with an independent regional retailer to design an automated analytics workflow to clean, model, and visualize transactional data.
- Surfaced key revenue drivers, discovering that premium customer segments generate 68% of total revenue.
- Mapped user lifecycles to identify a critical 40% decline in 3-month customer retention, delivering actionable data insights.
- BoolQ Question Answering Framework | PyTorch, Transformers, NLP, SVM
- Fine-tuned a BERT-base model on the open-source BoolQ dataset, achieving 87% accuracy in binary question-answering tasks.
- Engineered comparison feature pipelines using word2vec and fastText embeddings to benchmark transformer performance against traditional SVM baselines with TF-IDF.
- 🔗 GitHub Repository
- Enterprise MLOps Infrastructure Deployment | Flask, Docker, Celery, MLflow
- Containerized a Flask inference API using Docker to guarantee environment consistency across development and staging environments.
- Implemented Celery for asynchronous prediction task management, reducing API bottlenecks and minimizing user experience latency by 35% during simulated high-concurrency loads.
- 🔗 GitHub Repository
- Predictive Sports Analytics Engine | Python, LSTM, Scikit-learn
- Developed sequential time-series feature sets from historical athletic match logs to forecast sports outcomes.
- Evaluated LSTM network performance against traditional Logistic Regression and Random Forest models using AUC-ROC optimization and log-loss reduction metrics.
- 🔗 YouTube Demo
- Multilingual Health Assistant Chatbot | Python, NLP, Dialog Management
- Built an NLP-driven conversational agent capable of processing and translating health-related queries in 3 different languages.
- Implemented custom intent classification algorithms and Named Entity Recognition (NER) to seamlessly navigate over 50 distinct user paths across 10,000+ synthetic health query variants.
- 🔗 YouTube Demo
- Large-Scale Click Prediction Analytics Engine | Apache Spark, HDFS, Python
- Architected a distributed log analytics pipeline using Apache Spark RDD and DataFrame APIs to calculate click prediction traffic patterns over HDFS.
- 🔗 GitHub Repository
- IMDB Exploratory Data Analysis & Visualization | Python, Scikit-learn, Plotly
- Executed comprehensive exploratory data analysis (EDA) and built high-dimensional visualization pipelines using K-Neighbors and Decision Tree algorithms.
- 🔗 GitHub Repository
💼 CORPORATE EXPERIENCE
- Resident Engineer | PCI Prime Group (Escorts Kubota) Nov 2024 – Feb 2025
Faridabad, Haryana, India
- Engineered automated ETL pipelines using Python to ingest, clean, and analyze high-frequency IoT hardware metrics, cutting manual reporting latency by 15 hours monthly.
- Conducted statistical trend analysis on machinery output datasets to identify operational variance patterns, converting raw data into predictive metrics for maintenance scheduling.
- IT Coordinator | AIR INDIA (AISATS) Aug 2024 – Nov 2024
Gurugram, Haryana, India
- Performed exploratory data analysis (EDA) and text mining on 100+ incident logs to identify systemic failure patterns, implementing data-driven workflows that decreased incident recurrence by 20%.
- Structured raw troubleshooting data into standardized documentation schemas and knowledge-base architectures, reducing average resolution time for technical support teams.
- Associate Analyst, Content Engineering | GlobalLogic (Hitachi Group Company) Nov 2022 – Apr 2023
Gurugram, Haryana, India
- Partnered on data analytics operations utilizing built-in Python and ML tools for specialized computer vision applications, focusing on object recognition workflows.
- Processed high-fidelity source data pipelines to ensure exact model alignment, data integrity, and pipeline verification during early-stage testing.
🚀 CERTIFICATIONS
-
The Complete Machine Learning Course with Python | Udemy 2025
Core Focus: Statistics, Linear Algebra, EDA, Supervised/Unsupervised Learning, NLP, and Model Optimization.
-
Python 3: Project-based Python, Algorithms, Data Structures | Udemy 2025
Core Focus: Sorting Frameworks, Search Metrics, Graph Structures, Recursion, Backtracking, and Dynamic Programming.
-
Python 101 for Data Science | IBM 2024
🎓 EDUCATION
- M.Sc. in Data Science | HSE University 2021 – 2023
- Bachelor of Technology (B.Tech) | Harcourt Butler Technical University (HBTU) 2012 – 2017
📧 kanishkk202@gmail.com
🔗 WhatsApp
📱 M.No: +91-9990465116
🌍 Available for remote/office work, freelancer, mentoring & collaborations
📌 Languages Known