Atal Sharma

Data Scientist

Udhampur, Jammu & Kashmir, India7 yrs 8 mos experience

Most Likely To SwitchAI ML Practitioner

Key Highlights

Expert in developing end-to-end machine learning solutions.
Proficient in building scalable APIs and processing large datasets.
Achieved 90% accuracy in infant cry classification project.

Stackforce AI infers this person is a Data Scientist specializing in Healthcare AI applications and scalable machine learning solutions.

Contact

Skills

Core Skills

Large Language Models (llm)Application Programming Interfaces (api)Aws SagemakerMachine Learning

Other Skills

Machine Learning AlgorithmsPythonTensorFlow/KerasLibROSANumPySciPyMatplotlibScikit-learnPyDubAI AgentsFine TuningAgentic AI DevelopmentRetrieval-Augmented Generation (RAG)CassandraJupyter

About

Data Scientist with hands-on experience in developing and deploying end-to-end machine learning and deep learning solutions, including RAG-based LLM applications. Proficient in building scalable APIs and processing large-scale datasets using PySpark and other high-performance Python libraries, with a track record of delivering impactful data-driven projects.

Experience

Ingenia ai

Data Scientist

Jul 2023 – Present · 2 yrs 9 mos · Remote · Remote

Large Language Models (LLM)Application Programming Interfaces (API)

Warpspd.ai

2 roles

Data Scientist

Sep 2022 – Jul 2023 · 10 mos · Noida, Uttar Pradesh, India

Application Programming Interfaces (API)AWS SageMaker

Machine Learning Engineer

Sep 2020 – Sep 2022 · 2 yrs · Noida, Uttar Pradesh, India

Central university of jammu

M.Tech Scholar

Aug 2018 – Sep 2020 · 2 yrs 1 mo · Jammu Area, India · On-site

During my M.Tech program, I undertook an applied research project aimed at building an intelligent system to classify different types of infant cries using machine learning and deep learning techniques. The motivation behind this work was to assist pediatric care by enabling early identification of an infant’s needs based on crying patterns.
I personally collected high-quality infant crying audio samples from the children's section of a district hospital, following ethical guidelines and ensuring real-world data authenticity. The dataset was labeled into five cry types corresponding to different physiological or emotional states (e.g., hunger, pain, sleepiness, discomfort, and medical distress).
To develop the classification system, I applied Recurrent Neural Network (RNN)-based architectures, leveraging Python and several audio signal processing libraries including LibROSA, PyDub, and SciPy to extract relevant audio features such as MFCCs, pitch, energy, and spectrogram-based insights.
Key contributions of the project include:
End-to-end pipeline: From raw audio collection and preprocessing to model training and evaluation.
Custom feature engineering for better cry pattern recognition.
High model performance, achieving approximately 90% accuracy on the test dataset.
Experimented with various architectures and hyperparameters to optimize performance.
This work demonstrates the potential of deep learning in infant health diagnostics and paves the way for integrating AI in neonatal care environments.
Tools & Libraries Used: Python, TensorFlow/Keras, LibROSA, NumPy, SciPy, Matplotlib, Scikit-learn, PyDub.

Machine Learning AlgorithmsMachine Learning

Inrdeals private limited

Internship Trainee

Jul 2017 – Sep 2017 · 2 mos · Jammu, Jammu & Kashmir, India

Model institute of engineering and technology

Center for Research Innovation and Entrepreneurship

Jan 2017 – Apr 2018 · 1 yr 3 mos · Jammu, Jammu & Kashmir, India

Worked as an Intern under Center for Research Innovation and Entrepreneurship in Model Institute of Engineering and Technology.
Worked under patented project.