Abhinav Shandilya

AI Researcher

Gurugram, Haryana, India11 yrs 6 mos experience

Most Likely To SwitchAI Enabled

Key Highlights

11+ years of experience in machine learning solutions.
Expert in deploying Large Language Models at scale.
Proven track record of driving significant business impact.

Stackforce AI infers this person is a SaaS expert with a strong focus on machine learning and AI-driven solutions.

Contact

abhinavshandilya244@gmail.com LinkedIn

Skills

Core Skills

Large Language Models (llm)Deep LearningNatural Language Processing (nlp)Big Data AnalyticsExplainability In AiMachine LearningCausal InferenceDeep Reinforcement LearningGenerative AiBig Data

Other Skills

Data SciencePython (Programming Language)Apache SparkScientific ExperimentationA/B TestingMySQLData AnalysisProgrammingStatistical ModelingStatisticsExtract, Transform, Load (ETL)Artificial Intelligence (AI)SQLRTensorFlow

About

Experienced Applied Scientist / ML Engineer with 11+ years of building scalable machine learning solutions and driving business impact across multiple domains. I combine deep technical expertise with a strong research mindset, currently working on real-world applications of LLMs, agentic AI systems, and probabilistic modeling. I thrive at the intersection of applied ML and emerging AI research—passionate about designing intelligent systems that are robust, interpretable, and deployable at scale.

Experience

11 yrs 6 mos

Total Experience

1 yr 11 mos

Average Tenure

2 yrs 7 mos

Current Experience

Salesforce

Senior Machine Learning Engineer (SMTS)

Oct 2023 – Present · 2 yrs 7 mos · Gurugram, Haryana, India · Hybrid

Leading applied research and engineering efforts to deploy Large Language Models (LLMs) and agent-based systems in production. Focused on Retrieval-Augmented Generation (RAG), reinforcement learning from human feedback (RLHF), and explainability in LLM-powered QA systems. My work bridges foundational AI advances with real-world scalability, integrating state-of-the-art transformer models into Salesforce's Agentforce ecosystem.
Key Contributions:
Problem Record Classification: Designed and deployed a multi-task deep learning model for automated Problem Record classification, using sentence transformer embeddings to encode unstructured incident logs and task-specific softmax heads to predict blast radius, customer impact, and resolution categories. Enabled scalable semantic triage of production failures, improving classification accuracy and reducing manual analysis effort.
Log Anomaly Detection System:
Designed a transformer-LSTM hybrid integrated with a vector-store-based RAG pipeline, improving incident triage accuracy by 30% and reducing MTTR by 20% across enterprise support workflows.
LLM Fine-Tuning with RLHF:
Fine-tuned production-grade LLMs using human feedback loops to increase response relevance by 25%. Built reward modeling and feedback pipelines aligned with user satisfaction metrics.
Agentic Q&A System:
Developed a Slack-integrated autonomous agent leveraging Salesforce’s Agentforce framework, capable of ingesting structured (tables, metadata) and unstructured sources (PDFs, internal docs) to answer engineering queries in real-time.
Explainability and Trust in AI Systems:
Built an XAI layer for anomaly predictions and LLM outputs, delivering transparent, context-aware explanations to non-technical stakeholders, increasing trust and adoption in automated QA decisions.

Deep LearningData ScienceBig Data AnalyticsNatural Language Processing (NLP)Large Language Models (LLM)Python (Programming Language)

Uber

Applied Scientist II

Jul 2021 – Sep 2023 · 2 yrs 2 mos · Bangalore Urban, Karnataka, India · Hybrid

Drove a $10 million incremental annual revenue in uber mobility and enhanced marketing budget spend efficiency by 30%, utilizing a causal machine learning uplift model. The model computed customer-specific spend efficiency, and its real-world efficacy was validated through a meticulously designed A/B test using the explore-exploit technique
Improved the booking conversion rate by 5% through implementation of a vehicle recommendation system using a learning to rank framework for Uber car rentals
Designed and fine-tuned a comprehensive product funnel for Uber car rentals, encompassing instrumentation, KPI definitions, and effective dissemination strategies
Designed and analyzed various marketplace experiments by leveraging A/B testing, synthetic control and other quasi experimental design techniques to enable data driven decision making at Uber scale and marketplace complexities

Machine LearningCausal InferenceApache SparkDeep LearningScientific ExperimentationPython (Programming Language)+1

D cube analytics

Data Science Consultant

May 2020 – Jul 2021 · 1 yr 2 mos · Bengaluru, Karnataka, India · Remote

Elevated the engagement rate of 200K US HCP by 8%, leveraging a Deep Reinforcement Learning (DQN) based recommendation engine on AWS and MLFlow. Orchestrated the successful collaboration of an interdisciplinary team of 8 data engineers and data scientists
Successfully conceptualized and created a Knowledge Graph based on sales, insurance and marketing data of 200K HCPs in USA

Generative AIDeep Reinforcement LearningBig DataPython (Programming Language)A/B Testing

Rategain

Machine Learning Engineer

Mar 2018 – May 2020 · 2 yrs 2 mos · Noida, Uttar Pradesh, India · On-site

Led a team of four junior data scientist and ML engineers to raise price intelligence report acceptance by 5%. Automated quality checks with a deep-learning anomaly detection model, saving 300 man hours/month for a top travel firm
Elevated the average daily rate of a leading car rental company by 4% through the successful development and deployment of a dynamic price recommendation engine on Google Cloud Platform (GCP)
Enhanced the Net Promoter Score (NPS) of a prominent travel conglomerate by 3% through the creation of a topic modeling framework and sentiment analysis, leveraging BERT. Collaborated closely with product and engineering teams to translate insights into actionable strategies

Python (Programming Language)

Cartesian consulting

Data Scientist

Nov 2015 – Mar 2018 · 2 yrs 4 mos · Bengaluru, Karnataka, India · On-site

Achieved a 5% reduction in churn rate by devising key performance indicators (KPIs) to evaluate discount sensitivity and crafting a personalized offer and marketing campaign framework. Utilized metrics like CLTV, marketing affinity, customer segmentation, and discount responsiveness
Boosted the average basket size by 11% and elevated the NPS score by 5% through the development of a demand forecasting model for over 400 retail stores. The model harnessed customer demographic, transactional data, and store attributes to inform predictions
Drove a $250K increase in annual revenue by formulating and executing a merchandise allocation framework across diverse stores. Employed a constrained optimization approach to optimize allocation strategies

Python (Programming Language)

Mu-sigma business private limited

Trainee Decision Scientist

Sep 2014 – Oct 2015 · 1 yr 1 mo · Greater Bengaluru Area · On-site

Created a claim amount prediction model for automobile accidents, serving approximately 2 million subscribers of a prominent Australian insurance company
Empowered data-driven decision-making for pivotal stakeholders at a top Australian telecom operator. Engineered multiple KPIs and automated diverse Tableau dashboards

Python (Programming Language)