# Prajak Sen

Prajak Sen is a remote worker on Himalayas, Research Associate, Physical AI and World Models at BITS Pilani, based in India.

Source: https://himalayas.app/@prajaksen

```text
AI Forward Deployed Engineer at R Systems building an enterprise agentic sales platform with LangGraph, Salesforce and Microsoft Graph.
```

| Detail | Value |
| --- | --- |
| Availability | Looking for a job |
| Location | India |
| Authorized to work in | India, United States, United Kingdom, Spain |
| Salary expectations | 95k+ INR |
| Job categories | Forward Deployed AI Engineer, DevOps Engineer, Cloud Engineer |
| Skills | Agentic AI, Large Language Models, RAG, Multi agent orchestration, LLM Evaluation, Guardrails AI, Hybrid Retrieval, Dense Retrieval, Reranking, LLM Serving, Model Quantization, Machine Learning, Natural Language Processing, Explainable AI, Anomaly Detection |
| LinkedIn | https://linkedin.com/in/prajak-sen |
| GitHub | https://github.com/prajak002 |

## About

```text
At R Systems, I’m building an agentic sales-automation platform for a US enterprise in employee benefits, covering workflows from broker RFP intake through proposal and contract generation. I’m developing per-case LangGraph orchestration that combines model reasoning with deterministic validation.

I also drove the underwriting integration and aligned the architecture with client technology stakeholders. I defined the platform’s PHI security posture, including access controls, encryption and CI code scanning.

At Oro Labs, I own the Process Runner, a LangGraph engine for document-processing and procurement workflows. I also shipped a self-hosted Langfuse observability platform serving six teams and built an ingestion gateway that centralizes validation, quotas and tenant revocation.

My production AI work includes a vehicle-telemetry RAG system at Intangles Lab, which held sub-200 ms p95 latency, and quantized LLM serving that reduced per-token cost by 35%. Earlier, I worked on event-driven systems at Wayfair India, RAN anomaly detection at TEOCO Corporation, and cloud delivery automation at Red Hat.
```

## Experience

### Research Associate, Physical AI and World Models (current)

BITS Pilani. Jul 2026 - Present

```text
Working on multilingual instruction following for robotic action generation and dexterous retargeting on the ARCTIC dataset with MANO and SMPL-X simulation.
```

Skills: Robotics, VLA, MANO, SMPLX, PyTorch, Multilingual-NLP

### ML / AI Engineer (current)

Oro Labs. May 2026 - Present

```text
Owned the Oro Process Runner, a LangGraph orchestration engine for document-processing and procurement workflows. Architected a self-hosted Langfuse observability platform and proved resilience with a 13-scenario chaos and stress suite.
```

Skills: LangGraph, Langfuse, OpenTelemetry, Helm, GitOps, Spring-Boot, RAG, LLM-As-A-Judge, Chaos-Engineering, Kubernetes

### AI Forward Deployed Engineer (current)

R Systems

```text
Building an agentic sales-automation platform for a US enterprise client, covering the full sales lifecycle from broker RFP intake to contract generation. Designed stateless orchestration with LangGraph, integrated Salesforce and Microsoft Graph, and defined the PHI security posture.
```

Skills: LangGraph, Salesforce, Microsoft-Graph, Claude, OpenAI, LiteLLM, Kafka, Celery, RBAC, Microsoft-Entra-ID-(OAuth2-SSO), S3, KMS, Secrets-Manager, GitHub-Actions

### Visiting Researcher, Small Language Models

IISc Bengaluru. Mar 2026 - Jun 2026

```text
Studied 1B to 3B parameter models under hard compute budgets across INT8/INT4 quantization, structured sparsity, and KV-cache layout.
```

Skills: Small-Language-Models, Quantization, Sparsity, KV-Caching, PyTorch

### AI Scientist

Intangles Lab. Jan 2026 - Apr 2026

```text
Built a production RAG system for real-time vehicle-telemetry Q&A with sub-200 ms p95 latency. Designed metadata-aware chunking with hybrid retrieval, served quantized LLMs on vLLM, and cut unsafe outputs by 68% with a multi-layer guardrail sidecar.
```

Skills: Lang-Chain, LangGraph, Pinecone, Faiss, fastAPI, Amazon-EKS, VLLM, BM25, Reranking, GuardRails, Multi-Agent-Systems

### Software Engineer

Wayfair India. Jan 2025 - Nov 2025

```text
Designed Kafka event-driven pipelines for live inventory and order updates across suppliers, warehouses, and fulfilment. Hardened order-routing consumers to reduce failed and delayed events.
```

Skills: Kafka, Event-Driven-Architecture, Distributed-Systems, Java, SQL

### Research Intern, Explainability and Model Alignment

AryaXAI. Sep 2023 - Jan 2025

```text
Benchmarked SHAP, LIME, Integrated Gradients, and DLBacktrace on credit-risk models, with DLBacktrace scoring 31% higher fidelity. Shipped a FastAPI service for near real-time explanations under 300 ms.
```

Skills: SHAP, LIME, Integrated-Gradients, DLBacktrace, fastAPI, AutoGen, Langfuse

### ML Engineer, Radio Access Networks

TEOCO Corporation. May 2024 - Nov 2024

```text
Trained an XGBoost and LSTM ensemble on RAN telemetry for multi-class anomaly detection, cutting mean fault-response time by 30%. Built Spark streaming pipelines ingesting 1M+ log rows per hour.
```

Skills: XGBoost, LSTM, Spark, Anomaly-Detection, Telemetry, Python

### Software Engineer Intern, Cloud and DevOps

[Red Hat](https://himalayas.app/companies/red-hat). Jan 2024 - May 2024

```text
Automated microservice delivery with Argo CD GitOps, OpenShift, and Pulumi IaC, cutting release-cycle time by 50% and removing configuration drift across three environments.
```

Skills: ArgoCD, OpenShift, Pulumi, GitOps, Kubernetes, DevOps

### Research Intern

[Carnegie Mellon University](https://himalayas.app/companies/carnegie-mellon-university). Feb 2022 - Oct 2023

```text
Owned a 3D CNN pipeline in PyTorch for macromolecular structure recognition in cryo-electron tomograms, including simulated-to-real domain adaptation.
```

Skills: PyTorch, 3D-CNN, Domain-Adaptation, Cryo-ET

### Open Source Contributor

LitmusChaos. May 2023 - Aug 2023

```text
Wrote Go automation and Prometheus integrations for fault-injection experiments and triaged 30+ issues and pull requests.
```

Skills: Go, Prometheus, Chaos-Engineering, Open-Source

### Research Intern, Multilingual NLP

[IIT Roorkee](https://himalayas.app/companies/indian-institute-of-technology-roorkee). Nov 2022 - May 2023

```text
Built spelling normalization for code-mixed Hindi-English text and NER/POS pipelines on a 100k+ entry corpus, reaching ROUGE-L 0.917 and F1 0.933.
```

Skills: NLP, NER, POS, Code-Mixed-NLP, Python

## Education

### Institute of Engineering & Management, Kolkata

Bachelor of Technology, Computer Science and Engineering. 2020 - 2024. Grade: 9.16/10

```text
Pursued a Bachelor of Technology in Computer Science and Engineering, achieving a CGPA of 9.16 out of 10.
```

## Tech stack

- Google Cloud Platform
- Amazon S3
- Kubernetes
- Helm
- Amazon EKS
- CircleCI
- GitHub Actions
- Salesforce
- PostgreSQL
- Sentry
- Node.js
- Django
- Spring Boot
- Redis
- Terraform
- Pulumi
- React
- JavaScript
- Python
- Java
- Go
- C++
- PyTorch
- MLflow
- Kafka
- Celery
- FastAPI
- Grafana
- Prometheus
- OpenTelemetry
- SigNoz
- gRPC
- monday.com
- TypeScript
- Docker
- SQL
- XGBoost
- AWS KMS
- LangChain
- AutoGen
- Pinecone
- LiteLLM
- vLLM
- Langfuse
- Guardrails AI
- Argo CD
- ONNX Runtime
- Faiss
- LangGraph
- AWS
- Claude

To contact Prajak, message them from this profile with a Himalayas employer account: [Hire on Himalayas](https://himalayas.app/hire).

## People also viewed

- [Shivam Gulati](https://himalayas.app/@shivamgulati): Data Scientist, Quara AI Tech Pvt Ltd
- [Devansh Vora](https://himalayas.app/@devanshvora): Jr. AI Research Fellow, Smart Energy Learning Center Dhirubhai Ambani University
- [Krishna Yadav](https://himalayas.app/@krishnayadav4): AI Engineer (LangGraph), Elvz.ai
- [Rishabh Das](https://himalayas.app/@rishabhdas): Founding Engineer (AI/ML), Loylty Rewardz
- [Ayush Arya](https://himalayas.app/@ayusharya4): Member of Technical Staff 1, Oracle
- [Yash Rawal](https://himalayas.app/@yash_rawal): AI/ML Engineer, Bigscal Technologies Pvt. Ltd.
- [Onkar Kolekar](https://himalayas.app/@onkarkolekar): AI Engineer Intern, Pi labs
- [Parth Jain](https://himalayas.app/@parthjain): Research Intern, Connectomics and Computational Biology, PRAN Foundation
- [PREETI User](https://himalayas.app/@preetiuser9): AI Application Engineer II, Oracle
- [Preeti User](https://himalayas.app/@preetiuser10): AI Application Engineer II, Oracle
