Vaibhav Srivastava

Data Scientist

Ghaziabad, Uttar Pradesh, India4 yrs 9 mos experience

Highly StableAI Enabled

Key Highlights

Led architecture of enterprise semantic layer for data harmonization.
Engineered Generative AI solutions enhancing customer service efficiency.
Implemented advanced NLP systems for unstructured data insights.

Stackforce AI infers this person is a Data Science expert with a focus on Generative AI and NLP solutions.

Contact

Skills

Core Skills

Generative AiMachine LearningMlopsNatural Language Processing (nlp)

Other Skills

Gen AITensorFlowData ScienceGraph IngestionData ProcessingKnowledge GraphAPI DevelopmentPySparkAWSObject DetectionOpenCVCNNApache SparkData WranglingSpark architecture

About

As a Senior Data Scientist at ZS, I lead the architecture and implementation of an enterprise semantic layer to harmonize structured and unstructured data into canonical objects. My work includes developing GenAI feature-extraction pipelines and designing operational Copilot agent workflows for efficient knowledge-based querying. I also implement embedding and retrieval stacks and graph ingestion processes to support knowledge graph ingestion and advanced data analysis. With a B.Tech in Computer Science from Inderprastha Engineering College, I bring a strong foundation in machine learning and decision sciences. My core competencies include Generative AI, TensorFlow, and advanced data processing techniques, which I have applied to deliver impactful solutions across various industries. My focus remains on leveraging cutting-edge technologies to enhance data-driven decision-making and create scalable, innovative systems.

Experience

4 yrs 9 mos

Total Experience

3 yrs 4 mos

Average Tenure

1 yr 5 mos

Current Experience

Zs

Senior Data Scientist

Dec 2024 – Present · 1 yr 5 mos · Noida, Uttar Pradesh, India · Hybrid

Lead architect and implementer of an enterprise semantic layer that harmonizes structured (Redshift) and unstructured
(MSL notes, PDFs) sources into canonical objects: Document → Section → Statement → TextualEntity → OntologyTerm.
Built GenAI feature-extraction pipelines (NER + ontology mapping) that produce entity & statement records, embeddings, and
validated triples for knowledge graph ingestion.
Designed and operationalized a Copilot agent flow (planner/supervisor/chain of toolkits) allowing KBQ routing to deterministic
SQL, graph traversal, or RAG pipelines depending on KBQ type.
Implemented embedding & retrieval stacks (OpenSearch) and graph ingestion (triplet generation, SHACL validation,
Neptune/TTL loads) to support KBQ analytics and RAG.
Mentored engineers and coordinated with product owner and SMEs to turn KBQs into production use-cases and trusted assets.

Gen AITensorFlowNatural Language Processing (NLP)Data ScienceMachine LearningGraph Ingestion+1

Impetus

2 roles

Senior Machine Learning Engineer

Jul 2024 – Dec 2024 · 5 mos · Noida, Uttar Pradesh, India

Engineered and implemented a Generative AI solution for United Airlines, leveraging the Haiku LLM Bedrock foundation model to summarize customer PNR information, enhancing customer agent efficiency and service quality.
Optimized system performance by implementing asynchronous API calls and parallel processing, resulting in a 45% reduction in total response time and improved overall system efficiency.
Developed a comprehensive solution incorporating data processing, PII masking, real-time summary streaming, and an intuitive Streamlit-based UI, complemented by Grafana dashboards for statistical analysis and user feedback visualization.

MLOpsNatural Language Processing (NLP)Generative AIAPI Development

Machine Learning Engineer

Aug 2021 – Aug 2024 · 3 yrs · Noida, Uttar Pradesh, India

Data science:
Implemented NL-QA system on AWS to generate insights from unstructured claim reports using cutting-edge NLP models improving
efficiency and enabling 360-degree analysis.
Engineered pipeline leveraging semantic search, ranking, summarization, and metadata tagging to enable self-service analysis.
Developed and demonstrated a resume parser prototype to extract candidate details like skills, education, and experience into structured
JSON output using natural language processing techniques. It streamlines the process of understanding candidate profiles and aids in
making efficient project assignment decisions.
Built a customized video generator using Polly, SSML, and AI to inform passengers in multiple languages and accents.
Employed state-of-the-art NLP models like BERT to vectorize text and calculate semantic similarity for identifying textual parallels.

PySparkMachine LearningNatural Language Processing (NLP)