Looking for a job

Sahil Maheshwari

@sahilmaheshwari

Data Scientist, 12+ years. From modelling to team leadership across consulting, enterprise, and startup. Focused on LLM/GenAI and agentic AI.

India

Message

What I'm looking for

Director/VP/Head-level DS or AI leadership role where I can drive both technical direction and business outcomes. Core background in applied ML — pricing, recommendations, experimentation, platform engineering — with additional depth in LLM and GenAI. Open to consulting, enterprise, or well-structured startups.

I've worked across the full DS stack — from building models to architecting data platforms and MLOps infrastructure — across different environments: consulting (Happiest Minds, ThoughtWorks), enterprise (Cognizant, Sixt), and startup (Hunch).

At Hunch, I built the DS function from scratch: hired an 8-member team, shipped 10+ projects in three quarters, built recommendation systems, an LLM-powered content tagging and profiling layer, a scalable data platform on open-source tools (halving monthly costs), and an in-house MLOps platform that cut deployment time from days to minutes.

At Sixt, I worked on dynamic pricing using elasticity models (3% sustainable margin uplift), fleet planning simulations with ML and genetic algorithms, risk-based insurance pricing, and an internal A/B testing framework. At ThoughtWorks, I delivered pricing and search projects across retail and enterprise clients.

Earlier roles at Happiest Minds, Cognizant, and Infosys covered credit risk modelling, time series modeling, computer vision tasks, and dashboard building.

Since early 2025, I've been consulting independently — delivered a 4-month first-principles ML training programme for a fintech engineering team, published a technical article series on the geometric foundations of loss functions, and built LLM pre-training and fine-tuning pipelines from scratch. I also designed a framework-agnostic context engineering layer for agentic workflows, which has since evolved into an active open-source project.

My primary interests are LLM systems, agentic workflows, and ML platform engineering. I'm looking for leadership roles — Head of DS, Director, or VP — where both technical depth and business impact are expected.

Experience

Work history, roles, and key accomplishments

Current

Data Science Consultant

Current

Independent

Feb 2025 - Present (1 year 4 months)

- Designed and delivered a 4-month long first-principles ML training programme for an engineering team of a fintech client
- Published a series of research articles on the geometric foundations of loss functions
- Built LLM pre-training and fine-tuning pipelines from scratch, and designed a framework-agnostic context engineering layer for agentic workflows

First Principles Machine Learning Loss Function Research LLM AI Agent Frameworks Fraud Detection Supervised FineTuning PEFT

Head of Data Science

Hunch

Nov 2023 - Jan 2025 (1 year 2 months)

Built and led an 8-member data science team, delivering 10+ projects in three quarters. Developed recommendation models and LLM-powered content tagging/user profiling, architected a scalable open-source data platform that cut monthly costs by half, and built an in-house MLOps platform reducing model deployment time from days to minutes.

User Profiling Recommendation Systems Data Platform Architecture Team Leadership Cost Optimization LLM powered tagging MLOps Model deployment automation

Staff Data Scientist

Sixt

Oct 2020 - Aug 2023 (2 years 10 months)

Led dynamic pricing using elasticity models, delivering a 3% sustainable margin uplift. Built ML fleet planning via genetic algorithm optimization, implemented risk-based insurance pricing to balance conversion and loss ratio, and developed an in-house A/B testing framework; also led linear-programming fleet optimization during COVID for 10% annual cost savings.

Genetic Algorithms Fleet Planning Simulation Dynamic Pricing Elasticity modeling Insurance pricing optimization Expected loss modeling A B testing frameworks Linear programming

Lead Data Scientist

ThoughtWorks

Sep 2018 - Oct 2020 (2 years 1 month)

Drove consistent margin uplift across stores using hierarchical price-demand models. Designed a document search system using word2Vec, hierarchical clustering, and TF-IDF, significantly improving search conversion.

Information Retrieval Conversion Optimization Margin Optimization Hierarchical price demand modeling Document search Word2Vec Hierarchical clustering TF IDF

Senior Data Scientist

Happiest Minds

Oct 2015 - Sep 2018 (2 years 11 months)

Automated damage detection using transfer learning, moving from brief to production-ready model in weeks. Built real-time multi-feed CCTV tracking for store-level footfall analytics, developed risk-based pricing and time series forecasting pipelines, created an incremental education-grade prediction pipeline that avoided full retraining, and delivered R Shiny/Tableau dashboards with decision-ready insights.

Transfer Learning Real Time Tracking Time series Forecasting Claim propensity modeling Incremental prediction pipelines R Shiny Tableau

Product Specialist

Cognizant

Jun 2014 - Oct 2015 (1 year 4 months)

Led capability building and business development for the credit risk practice in partnership with FICO. Replaced rule-based credit scorecards with ML underwriting models, improving risk discrimination (precision/recall) and enabling more granular cutoff strategies.

Credit Scoring Precision Recall Optimization ML underwriting Cutoff strategy optimization FICO partnership Credit Risk Modeling Feature Strategy Business Development

Software Engineer

Infosys

Jan 2010 - Apr 2012 (2 years 3 months)

Delivered projects using Java and SQL, including automated testing with QTP. Built an employee summary dashboard improving visibility into resource utilization and allocation efficiency.

Dashboard Development Java SQL QTP Resource utilization analytics