Building production AI systems, agentic platforms, and large-scale ML infrastructure for finance, AdTech, analytics, and cloud products.
AI Architect with 10+ years defining enterprise AI strategy and scaling production AI platforms across finance, adtech, analytics, and cloud. Expert in GenAI, Agentic AI, LLM infrastructure, distributed inference, and data-to-AI platform engineering. Proven record of establishing architectural standards, driving org-wide platform adoption, and translating executive vision into autonomous AI capabilities. Strategic technical leader who aligns cross-functional stakeholders, governs architecture decisions, mentors engineering teams, and champions enterprise-scale AI transformation.
Currently architected the enterprise AI architecture strategy at AngelOne for a platform serving 40M+ retail investors (10M DAU), leading the design of an org-wide Agentic AI Platform processing 1M+ daily inferences. Previously built real-time NLP and experimentation infrastructure at Goldman Sachs and Google Cloud, and led NLP2SQL and AdTech optimization platforms at Slintel and Amagi.
Technical leader with deep experience in AI architecture, platform ownership, and cross-functional execution. Drives AI adoption through mentoring engineers, defining architecture standards, and leading technical initiatives from concept to production. Proven ability to own end-to-end AI systems — from model selection and inference optimization to observability, guardrails, and deployment.
| Impact | Context |
|---|---|
| 35% improvement in customer engagement | Multi-agent wealth management platform at AngelOne |
| 60% reduction in advisor response latency | Autonomous portfolio intelligence and investment advisory |
| 30% improvement in CTR | Real-time AdTech optimization at Amagi |
| 20% increase in viewer retention | Dynamic ad inventory optimization with edge ML |
| 200% improvement in dashboard creation efficiency | NLP2SQL platform at Slintel |
| 3x reduction in dashboard load times | Automated aggregation framework |
| 25% reduction in code-induced failures | Self-healing experimentation infrastructure at Google |
| 7% boost in long-term retention | ML-based A/B test orchestration at Google |
| 10x reduction in processing latency | ML pipeline migration at Persistent Systems |
AngelOne · Architect · Jan 2024 – Present Defined the enterprise AI architecture strategy for a platform serving 40M+ retail investors (10M DAU), leading the design of an org-wide Agentic AI Platform processing 1M+ daily inferences with standards for agent orchestration, memory, RAG, and tool integration. Drove cross-functional alignment across product, compliance, and engineering to deliver first-to-market AI Portfolio Intelligence, Auto Rebalance, and RM Enablement for 500+ RMs, improving engagement by 35%, reducing advisor latency by 60%, and cutting onboarding time by 50%.
Amagi · Tech Lead · May 2023 – Jan 2024 Owned the architecture of a distributed AI platform serving 1B+ impressions/day across 50+ streaming services at <50ms p99. Defined the vision for edge inference (ad personalization) and centralized LLM services (content classification, campaign optimization), delivering 30% CTR lift and 20% retention improvement.
Slinthead · Lead SWE · Oct 2021 – Mar 2023 Drove the architectural vision for an AI analytics platform serving 5K+ enterprise users and 100K+ queries/day, transforming BI consumers into self-service analysts through NLP2SQL and intelligent query optimization. Defined the LLM fine-tuning and aggregation architecture, improving dashboard creation efficiency by 200%, reducing load times by 3x, and displacing manual workflows to accelerate time-to-insight.
Google Cloud · Sr. SWE · Dec 2019 – Sep 2021 Designed an AI-powered software delivery platform adopted by 50+ teams processing 1000+ deployments/day, integrating CodeBERT-based code intelligence and deployment telemetry to predict release risk and recommend canary strategies. Established ML pipeline architecture leveraging code changes, rollbacks, and production failures as training signals, reducing code-induced failures by 25% and enabling safe forward deployment at scale.
Goldman Sachs · Sr. SWE · Aug 2018 – Dec 2019 Designed an AI-assisted client communication platform serving 5K+ bankers and processing 500K+ emails/day, leveraging BERT fine-tuned on enterprise and SME-curated datasets. Enabled contextual response recommendations, semantic search, and intelligent prioritization, reducing response latency and improving service consistency at scale. Defined automated retraining and phased deployment pipelines ensuring zero-downtime model updates.
Persistent Systems · SWE · Jul 2013 – Aug 2018 Designed predictive workforce optimization serving 10K+ field agents and 50K+ routes/day for task forecasting and resource allocation, improving productivity and reducing downtime. Led migration of legacy SAS workloads to scalable PySpark ML pipelines, achieving 10x latency reduction while eliminating licensing costs.
LangGraph · LangSmith · Guardrails · MCP · A2A · vLLM · Hugging Face · Unsloth · Claude · DSpy · Context/Harness Engineering · Agent Orchestration/Governance/Evaluation · Memory Systems · AI-Assisted SDLC · Vector DB
PyTorch · MLflow · Ray · FastAPI · Vertex AI · Bedrock · AI Foundry · AgentCore
Spark · Flink · Kafka · Qdrant · PostgreSQL · Databricks
AWS · GCP · Azure · Docker · Kubernetes · Terraform · Airflow · CI/CD
Python · SQL · Bash · Java · Go · Rust
AngelOne · 2024–Present
Enterprise AI architecture for a platform serving 40M+ retail investors (10M DAU), designing an org-wide Agentic AI Platform processing 1M+ daily inferences with standards for agent orchestration, memory, RAG, and tool integration. Delivered first-to-market AI Portfolio Intelligence, Auto Rebalance, and RM Enablement for 500+ RMs, improving engagement by 35%, reducing advisor latency by 60%, and cutting onboarding time by 50%.
Amagi · 2023–2024
Distributed AI platform serving 1B+ impressions/day across 50+ streaming services at <50ms p99. Defined the vision for edge inference (ad personalization) and centralized LLM services (content classification, campaign optimization), delivering 30% CTR lift and 20% retention improvement.
Slinthead · 2021–2023
AI analytics platform serving 5K+ enterprise users and 100K+ queries/day, transforming BI consumers into self-service analysts through NLP2SQL and intelligent query optimization. Defined the LLM fine-tuning and aggregation architecture, improving dashboard creation efficiency by 200%, reducing load times by 3x, and displacing manual workflows to accelerate time-to-insight.
Google Cloud · 2019–2021
AI-powered software delivery platform adopted by 50+ teams processing 1000+ deployments/day, integrating CodeBERT-based code intelligence and deployment telemetry to predict release risk and recommend canary strategies. Established ML pipeline architecture leveraging code changes, rollbacks, and production failures as training signals, reducing code-induced failures by 25% and enabling safe forward deployment at scale.
Goldman Sachs · 2018–2019
AI-assisted client communication platform serving 5K+ bankers and processing 500K+ emails/day, leveraging BERT fine-tuned on enterprise and SME-curated datasets. Enabled contextual response recommendations, semantic search, and intelligent prioritization, reducing response latency and improving service consistency at scale. Defined automated retraining and phased deployment pipelines ensuring zero-downtime model updates.
Persistent Systems · 2013–2018
Predictive workforce optimization serving 10K+ field agents and 50K+ routes/day for task forecasting and resource allocation, improving productivity and reducing downtime. Led migration of legacy SAS workloads to scalable PySpark ML pipelines, achieving 10x latency reduction while eliminating licensing costs.
Production-grade LangGraph templates for enterprise agentic systems — multi-agent orchestration, hierarchical graphs, supervisor patterns, and RAG platforms. Open-sourced for community adoption.
- Mentored 30+ junior engineers across multiple organizations
- Conducted 100+ technical interviews
- Led and managed a team of 6 engineers
- Architecture owner for AI platform initiatives
- Frequent speaker and mentor on GenAI, NLP, and distributed ML systems
- Established engineering standards: type safety, test coverage, structured logging, documentation-first
M.Tech, Computer Science · IIT Bombay · 2015–2017
B.E., Computer Science · Nagpur University · 2009–2013
GATE 2015 · 99.51 percentile (among 115,425 registered students)
- Established architectural governance processes, design review boards, and technical standards across AI teams
- Mentored 30+ engineers across levels, conducted 100+ interviews, and led a team of 6 engineers
- Speaker and thought leader on GenAI, Agentic AI, LLM infrastructure, and enterprise AI adoption
- Defined production reliability standards for AI systems, SLOs, incident management, and post-mortems
- Published open-source reference architectures and platform blueprints for enterprise-grade agentic AI systems
- Scored 99.51 percentile in GATE 2015 among 115,425 registered students
LinkedIn · GitHub · sushant.ai · sam1064max@gmail.com · +91-8149040621