Sriram G

Spearheaded 8+ Generative AI projects across industries, driving a 30% increase in project delivery success.

• Increased project success rate by 30% across GenAI initiatives.
• Mentored 15 team members, leading to a 25% promotion rate.
• Executed 5 GenAI workshops, impacting 200 professionals by creating COCs and by the creation of Byte Size Programs.
• Led the team to deliver solutions, improving client satisfaction by 25%.
• Established a Center of Competency (CoC) for GenAI and NLP, which helped in the upskilling of Data Scientists on current GenAI development.
• Delivered end-to-end LLM-based solutions, including prompt engineering, RAG pipelines, and multi-agent orchestration.
• Collaborated with business, engineering, and product teams to define success metrics, and scale solutions in production.
• Led strategic engagements from ideation to deployment, incorporating Responsible AI principles into risk mitigation, and governance frameworks.
• Led the efforts in the creation and launch of Byte-Sized learning programs.
• Fostered an innovation culture aligned with growth mindset principles.

PROJECTS:

1. Index Forecasting and Dynamic Pricing

GenAI Powered Index Forecasting and Dynamic Pricing solution integrating 200+ dynamic signals, including macroeconomic indicators, global commodity trends, and inventory levels.

Achieved 98–99% forecast accuracy, driving an anticipated revenue uplift of USD 1 million through improved pricing and hedging decisions.

2. QueryBoT

Build an Q&A model leveraging a blended approach of using Azure Open AI models and building custom models basis the client requirement

• Handling the ambiguous acronyms and hallucinations.

• Deployment of the model to support near-real-time response from the system.

• Reduction in response time by 75%

3. Product De-duplication

Deduplicate 40 different sources to create product golden records at the parent organization level, and fine-tune extractive LLM models.

• Designed an extractive Q&A model to extract product attributes from non-semantic product descriptions.

Deployment of the model to support offline batch processing on a monthly basis to augment the financial reporting system.

• 40% reduction in manual effort in creating golden records

• Designed and delivered scalable AI/ML solutions for insurance clients; applied NLP, CV, and voice-related techniques on unstructured data, such as PDFs, images, and call transcripts.
• Led RFPs, PoCs, and solutioning, contributing significantly to business development.
• Collaborated with global, cross-functional teams on model deployment and model management.
• Pioneered process automation and optimization for enterprise workflows.

Projects:

1. Roof Top Detection:

Aerial image classification, object detection, and building an end-to-end MLOps pipeline.

• 30% reduction in manual roof reviews by adjusters and underwriters.

2. Prospect Analytics- Abandoned Quotes

Quote conversion model; real-time scoring of 0.5M quotes/day

• 26% increase in conversion; increased agent performance, and lead prioritization.

3. Acord Form Automation:

CV and OCR solution techniques to automate quote creation from PDF and scanned forms.

• 30% enablement of straight-through processing of Acord form automation

4. AI-based Ticket Assignment and Resolution

NLP-based ticket routing and resolution similarity matching.

• 15% reduction in the reallocation of tickets.

5. Aadhar Masking

CV pipeline to classify & redact Aadhar info in PDFs

Reduction of team size from 15 to 2 members.

• Acted as SME for data science projects, converting business problems into AI use cases.
• Delivered innovative solutions using NLP models and advanced analytics to build Extractive Text Summarization.
• Liaise with Subject Matter Experts, technical folks for building effective data science solutions.
• Mentor and provide guidance to team.

• Built ETL pipelines and BI dashboards for a Fortune 100 retail client; conducted root cause analytics and optimized KPIs.