Ankit Mani

Data Scientist

Bengaluru, Karnataka, India8 yrs experience

AI EnabledAI ML Practitioner

Key Highlights

7+ years of experience in AI and ML solutions.
Generated multi-million dollar business impact.
Expertise in NLP and predictive modeling.

Stackforce AI infers this person is a Data Scientist specializing in AI and ML solutions for SaaS and Beverage industries.

Contact

Skills

Core Skills

Natural Language Processing (nlp)Machine LearningTime Series ForecastingData Analysis

Other Skills

ARIMAAWS SageMakerAirflowAlgorithmAzure DevOpsBERT (Language Model)BERTopicClassificationCluster AnalysisDeep LearningGenerative AILLM-based embeddingsLarge Language Models (LLM)NLP pipelinesNeural Networks

About

Data Scientist with 7+ years of experience delivering end-to-end AI, ML, and NLP solutions, with deep expertise in building scalable AI-powered platforms using RAG, LLMs, embeddings, and custom models. I specialize in turning complex data into actionable insights, driving intelligent automation, and enabling predictive decision-making at scale. My technical toolkit includes Python, Scikit-learn, TensorFlow, SQL, Tableau, and AWS SageMaker, with hands-on experience in Deep Learning, Neural Networks, and time-series forecasting. I’ve built and fine-tuned LLM-based systems using prompt engineering and embeddings, applying advanced NLP techniques such as BERTopic, TF-IDF, sentiment analysis, topic modeling, and clustering. Proficient with Airflow and real-time pipelines, I bring a strong foundation in statistics, regression, classification, and predictive modeling. With a proven track record of generating multi-million dollar business impact, I thrive at the intersection of advanced analytics and real-world execution, collaborating with global stakeholders to solve high-value problems.

Experience

Salesforce

Data Scientist

Jul 2024 – Present · 1 yr 8 mos · Bengaluru, Karnataka, India · Hybrid

◙ RAG based Knowledge Retrieval Model for Agentforce:
Developed a RAG model to optimize content selection for user queries, uncovering gaps in Agentforce knowledge base and identifying underrepresented topics. Improved Agentforce’s retrieval efficacy by 7% through enhanced response coverage and content alignment.
◙ AI Analytics Platform for Agentforce :
Designed and deployed end-to-end NLP pipelines with LLM-based embeddings and custom models (relevance, sentiment, topic, Top-N utterances) to analyze 200K+ support chats monthly, driving $4M+ annual savings through intelligent case deflection and proactive support optimization.
◙ Top Question Driver (TQD) :
Developed an AI model using TF-IDF and BERTopic to
identify top user-discussed questions on the forum, enabling the business to take
proactive measures against emerging issues and reducing ticket inflow by 1%.

Natural Language Processing (NLP)Word EmbeddingsBERT (Language Model)Large Language Models (LLM)Generative AIPrompt Engineering+5

Ab inbev

Senior Data Scientist

Feb 2022 – Jun 2024 · 2 yrs 4 mos · Bengaluru, Karnataka, India · Hybrid

◙ Empties return forecast : Developed a forecasting model, from initial prototyping to full-scale deployment, incorporating advanced forecasting techniques and time-series stabilization methods. Achieved an 88% accuracy for short-term forecasts and 82% accuracy for long-term forecasts in diverse zones, resulting in decreased production losses and container procurement costs, leading to savings of $3.5 million.
◙ Fuel consumption: Developed a fuel consumption prediction model for fleet vehicles, resulting in a significant efficiency increase from 1.24 km/l to 1.72 km/l and generating savings of $1.2 million. Also, provided safety insights at the driver, vehicle, and route levels, and introduced a driver scoring system.
◙ Capacity Planning: Executed strategic reporting and budget analysis to optimize CapEx, utilizing Azure DevOps for an automated CI/CD pipeline. The reporting tool accurately assessed current and future asset needs, and identified CapEx fluctuations, leading to a 2% reduction by addressing gaps.
◙ Collaborated with business stakeholders to develop and implement methodologies for quantifying value across multiple projects. Additionally, actively contributed to the planning of projects for the upcoming year.

pythonDeep LearningLarge Language Models (LLM)Machine LearningGenerative AITime Series Forecasting+4

Philips

Data Scientist

May 2021 – Feb 2022 · 9 mos · Bangalore Urban, Karnataka, India

◙ Analysis of the phishing campaign and mandatory training that assisted in the identification of offender profiles, led to a reduction in the click-through rate from 12% to 9%.
◙ Developed 3 POCs and operational reports for various organizational metrics and indicators.

pythonMachine Learning

Atos

2 roles

Data Scientist

Promoted

Jan 2019 – May 2021 · 2 yrs 4 mos

◙ Designing and implementation of IT support tickets categorization system for automatic classification which helped to reduce the response SLA by 40%.
◙ Applied topic modeling and sentiment analysis techniques to extract text aspects and the sentiment expressed by customers.
◙ Developed an algorithm to estimate the incident volume of the IT Support Ticket for the next 30 days.

pythonMachine LearningClassificationNatural Language Processing (NLP)

Data Analyst

Feb 2018 – Jan 2019 · 11 mos

◙ Developed a dashboard to categorize and analyze call performance, assess device performance, and identify the best and worst-performing devices, which helped reduce the call drops by 30%.
◙ Developed scripts to gather data from Service-now, perform analysis, and provide insights using Python.

pythonStatisticsData Analysis