Soman Y.

Data Scientist

Gurugram, Haryana, India3 yrs 9 mos experience

Most Likely To SwitchAI Enabled

Key Highlights

Expert in AI governance and ethical frameworks.
Proven track record in model risk management.
Skilled in developing Generative AI solutions.

Stackforce AI infers this person is a skilled AI Governance and Data Science professional with a focus on ethical AI solutions.

Contact

Skills

Core Skills

Model GovernanceData AnalysisAi GovernanceGenerative AiData ScienceMachine LearningWeb Development

Other Skills

AWSAlgorithmsAnalytical SkillsAnalyticsAnyLogicArtificial Intelligence (AI)AutomationBayesian Linear RegressionBusiness AnalysisCSSCommunicationCompliance MetricsDashboardDashboard CreationData Analytics

About

Over 2.5 years scaling AI governance to protect models from ethical/risk threats. I build AI systems that don't break, don't hallucinate, and don't get the company sued.How can we ensure the transformative power of AI is harnessed for good? That's the question that fuels my work in AI Governance. This challenge drives my work in AI Governance, where I transform complex technical concepts into practical, ethical frameworks that organisations can actually implement. During my tenure at Thomson Reuters, I led the development of AI governance frameworks. Working alongside data scientists and policymakers, I bridged the gap between technical innovation and responsible implementation, reducing potential AI risk factors. I believe that life's greatest moments and deepest connections are outside our comfort zones. I'm driven to embrace challenges, learn continuously, and contribute to a future where AI is a force for good in the world.Let's connect!

Experience

American express

Analyst - Data Science (Model Risk Management)

Sep 2024 – Present · 1 yr 6 mos · Gurugram

Model Risk Management Group - Strategic Initiatives and Model Governance
➤ Strengthened enterprise model governance by enhancing the Model Impact Classification for 450+ models, improving accuracy and prioritization of risk decisions
➤ Improved validation quality by redesigning the tollgate process using scored checklists and raising first pass approvals from 58 percent to 86 percent in six months
➤ Delivered high-value automation including bulk validation report extraction for 1000+ reports, automated Performance Tracking Report downloads, and a completeness tracker that eliminated manual work and strengthened documentation control
➤ Supported key regulatory deliverables by implementing the Control and Compliance metrics framework and performing an OSFI E-23 policy gap analysis to ensure alignment with upcoming supervisory expectations

Model Risk ManagementData GovernanceAutomationCompliance MetricsModel GovernanceData Analysis

Thomson reuters

AI/ML Model Steward

Jul 2023 – Sep 2024 · 1 yr 2 mos · Bengaluru, Karnataka, India

Worked in a multifaceted role with the Thomson Reuters Data & Model Governance (Responsible AI) global team, alongside SurePrep LLC, part of Thomson Reuters.
➤ Collaborated with a global Model Governance team across four countries, enhancing AI ethical compliance and boosting model review completion from 35% to 95%.
➤ End-to-End in-house feature development of conversational AI using LangChain, Snowflake and Large Language models by designing, building, and executing highly customised Generative AI solutions handling information of all the models of TR.
➤ Led Model Governance for Thomson Reuters AI models, ensuring ethical compliance, risk assessment, robust documentation and managing Thomson Reuters model registries while educating stakeholders on model governance principles and aligning stakeholder roles within the RASCI framework for transparent AI/ML model lifecycle management.
➤ Foundational and Applied R&D : Leveraging LangChain, LlamaIndex, AWS Bedrock, Azure OpenAI and Large Language Models (including GPT-4, Gemini, Claude, Llama, etc.) for custom-tailored Generative AI production solutions with RAG architecture in Document Classification, Optical Character Recognition, Document QnA, and Document Layout Analysis.
➤ Focused on agile product development of production-ready Generative AI solutions with LLMs, RAG, MuRAG architectures, including successful Proof of Concept (PoC) implementations.
➤ Developed Agent-driven Generative AI solutions with the application of vector database systems such as AWS Kendra, Milvus, FAISS etc.
➤ Ensured model Explainability and Interpretability (Explainable AI).

LangChainAWSLarge Language ModelsGenerative AIData GovernanceSnowflake+1

Genpact

Data Scientist

Jan 2023 – Jun 2023 · 5 mos · Bengaluru, Karnataka, India · On-site

Worked with the Data Science and Insights team (AI/ML practice at Genpact).
➤ Developed market mix models and built simulations using AnyLogic.
➤ Utilized Bayesian linear regression with Python (PyMC3 library) to solve real-world problems.
➤ Researched and implemented Bass diffusion models for supply chain and product diffusion simulations.
➤ Expanded knowledge in optimization and solved Travelling Salesman Problem using PuLp CBC Solver and Google OR-Tools Solver.
➤ Worked on vehicle routing problem optimization problems, considering factors like the number of vehicles, delivering hubs, and demand.
➤ Applied expertise in various domains, including market mix models, Bayesian linear regression, and Bass diffusion models.
Skills: AnyLogic, Python, Java, Market Mix Models, Bayesian Linear Regression, Bass Diffusion Models, Optimization (Travelling Salesman Problem, Vehicle Routing Problem), Machine Learning

AnyLogicPythonBayesian Linear RegressionOptimizationData Science

Amazon

Mentee - Amazon ML Summer School

Jun 2022 – Jul 2022 · 1 mo

➤ Executed a comprehensive project on Time Series Forecasting, encompassing data analysis, machine learning, and dashboard creation.
➤ Employed Python libraries including Statsmodels, Sklearn, and Sktime to implement and deploy 9 distinct ML models.
➤ Developed a dynamic web application capable of accurately predicting future stock values for various countries, leveraging the trained ML models.
➤ Integrated a database system and utilized Streamlit to architect an interactive and intuitive dashboard, enhancing user experience and accessibility.
Skills: Python, Supervised Learning, Unsupervised Learning, Dimensionality Reduction, Causal Inference, Reinforcement Learning

PythonMachine LearningDashboard Creation

Google developers

Google DSC Lead

Jul 2021 – Jul 2022 · 1 yr

Worked as a Google DSC Lead, leading & managing community of 500+ students in GDSC NIIT community.
➤ Appointed as the Lead by Google in a highly selective process. 352 Leaders were selected from across India.
➤ Delegated responsibilities, organized events, and got in touch with industry experts to share the experience with the community.