Disha Maheshwari

Data Scientist

Bengaluru, Karnataka, India3 yrs 9 mos experience

Key Highlights

Expert in designing machine learning models for business impact.
Proven track record in video analytics and anomaly detection.
Strong background in data analysis and statistical modeling.

Stackforce AI infers this person is a Data Scientist skilled in AI and machine learning across various industries.

Contact

Skills

Core Skills

Deep LearningAnomaly DetectionNlpSentiment AnalysisMachine LearningBusiness AcumenData AnalysisStatistical Modeling

Other Skills

CNNCommunicationComputer VisionCritical ThinkingFinetuning LLMGRUGoogle BigQueryKerasLSTMLarge Language Models (LLM)Market AnalysisMicrosoft ExcelMicrosoft Power BIMicrosoft PowerPointPU Learning

About

With a passion for transforming complex data into actionable strategies, Disha is an experienced Data Scientist adept at navigating intricate datasets to uncover compelling narratives and drive tangible business outcomes. Her approach combines deep technical proficiency with a keen understanding of business operations, ensuring that every analysis and model developed is directly aligned with strategic objectives. Tech Stack: Python (Numpy, Pandas, Scikit-learn, TensorFlow, PyTorch), SQL, Git, Docker Machine Learning & AI: NLP, Transformers, Generative AI (LLMs, Prompt Eng., RAG, Fine-tuning), Hugging Face, LangChain Cloud & Databases: AWS, Azure, Vector Databases (FAISS, Pinecone, Chroma)

Experience

3 yrs 9 mos

Total Experience

1 yr 3 mos

Average Tenure

6 mos

Current Experience

Intellithink industrial iot

Data Scientist

Oct 2025 – Present · 6 mos · Bengaluru, Karnataka, India

Upwork

Data Scientist

Oct 2024 – Oct 2025 · 1 yr

Designed and implemented a CNN-LSTM pipeline for CCTV anomaly detection, achieving 80% accuracy and enabling automated real-time surveillance alerts.
Built an end-to-end prototype, applying data augmentation (fog and haze) to enhance model robustness and ensure reliable video analytics performance across diverse environmental conditions

Deep LearningAnomaly DetectionPythonNLPTensorFlowKeras

Freelancer.com

Data Scientist

Jun 2024 – Jul 2025 · 1 yr 1 mo · Remote

Built a No-Reference Video Quality Assessment (NRVQA) pipeline for video data, leveraging spatiotemporal and perceptual features, and ensemble models (XGBoost, LightGBM, Neural Nets) to achieve an RMSE of 0.774.
Developed a multiclass sentiment analysis system for post-launch product feedback, applying Bi-GRU and Bi-LSTM networks on large-scale Twitter data to classify across five sentiment categories and deliver actionable insights.

PythonNLPGRULSTMSentiment AnalysisStreamlit

Zs

Data Science Associate

Jun 2022 – May 2024 · 1 yr 11 mos · Bengaluru, Karnataka, India · Hybrid

Designed and deployed an XGBoost-based sales intelligence ecosystem for Microsoft product suites, processing 10M+ annual leads and boosting qualified lead conversions by 5–7%, with end-to-end ownership of pipeline (model training, deployment, and Power BI dashboards) and continuous MLOps monitoring across three production models.
Developed and integrated a semi-supervised ML model (PU Learning) on medical claims data for a major pharma client, improving HCP targeting accuracy and driving a 10–15% uplift in revenue potential, while ensuring transparency through SHAP-based model interpretability.

Machine LearningProblem SolvingBusiness AcumenCommunicationStakeholder Management