Open to AI Engineer roles

Hi, I'm Suchit Raj

AI Engineer // agentic LLM systems

I build agentic LLM systems — LangGraph, RAG, semantic search — on a foundation of Python, statistics, and ML, with production experience in enterprise integrations, AWS migration, and QA testing for the Commonwealth of Massachusetts.

View projects Get in touch
agent_trace.log running
suchit@agent ~ $ research "How does hybrid search improve RAG recall?"
planner decompose question → sub-queries
retriever BM25 + pgvector · cross-encoder rerank
synthesizer draft answer with citations
critic faithfulness check … pass
✓ cited answer returned · trace → Langfuse
365
external partner hosts validated for a MOVEit MFT upgrade
10,000+
customer records modeled for churn prediction
4-node
agent graph: planner, retriever, synthesizer, critic
EC2 → RDS
SQL Server backend migration supported on AWS
01

About

Portrait of Suchit Raj
Boston, MA B.S. Information Technology

I'm an Information Technology graduate from the University of Massachusetts Boston, focused on AI engineering: agent graphs that plan and call tools, retrieval pipelines that actually find the right passage, and evaluation harnesses that tell me whether a change made things better.

Originally from India, I moved to the United States at 15 and studied across Vermont, New York, and Colorado before landing in Boston. That path made me adaptable and quick to pick up new systems — which is most of the job.

Alongside the AI work, I've shipped in production settings: validating hundreds of partner integrations, supporting an AWS database migration, and writing QA test plans for the Commonwealth of Massachusetts.

University of Massachusetts Boston

May 2026

B.S., Information Technology

Relevant coursework Python Programming · Machine Learning & Model Training · Enterprise Software · Project Management (MS Project)

02

Experience

Integrations Intern

Commonwealth of Massachusetts — EOTSS

Jul 2026 — Present
  • Validated connectivity for 365 external partner hosts during a MOVEit MFT platform upgrade, running per-host transfer tests and telnet port checks via Remote Desktop Connection Manager to confirm endpoint reachability ahead of cutover.
  • Triaged failed connections by identifying firewall and ACL blocks, then coordinated directly with external customer contacts to whitelist Commonwealth source IPs — restoring transfers without escalation.
  • Supported migration of the MOVEit SQL Server backend from self-managed EC2 to Amazon RDS, alongside an OS upgrade from Windows Server 2016 to 2025.
  • Designed and executed QA test plans covering file transfer workflows, scheduled automation tasks, and user access paths; documented defects, verified fixes, and reported results prior to sign-off.
  • Presented migration scope, testing methodology, and results at the agency intern showcase to technical and non-technical stakeholders.
MOVEit MFTAWS EC2Amazon RDSSQL ServerQA TestingNetworking

System Assistant

Sodexo — University Dining

Sep 2023 — May 2026
  • Maintained enterprise POS systems under peak load, resolving real-time exceptions, managing access control, and providing cross-functional technical support.
POS SystemsAccess ControlTechnical Support
03

Projects

Featured · Agentic AI

Agentic Deep Research Assistant

LangGraph, RAG & Evaluation

01Plannerdecomposes the question
02Retrieverweb search · document retrieval
03Synthesizerwrites the cited answer
04Criticchecks the draft
  • Built a multi-agent research system in LangGraph (planner, retriever, synthesizer, and critic nodes) that decomposes questions, calls web-search and document-retrieval tools, and returns cited answers; enforced typed tool inputs/outputs with Pydantic AI and routed models through LiteLLM.
  • Implemented a textual RAG pipeline on Postgres/pgvector with HuggingFace sentence-transformer embeddings, hybrid BM25 + vector semantic search, and cross-encoder reranking.
  • Created an evaluation harness (golden question set, retrieval recall@k, LLM-as-judge faithfulness scoring) with traces in Langfuse, using results to iterate on prompts and agent graph design.
  • Served via FastAPI + Streamlit in Docker on AWS (EC2, S3).
LangGraphPydantic AILiteLLMPostgres/pgvectorHuggingFaceBM25LangfuseFastAPIStreamlitDockerAWS
View on GitHub
LLM · Support Agent

Client History Lookup Chatbot

LLM-Powered Jira Support Agent

question → intent classifier → Jira query → summary

  • Developed an LLM-powered support agent using Python, FastAPI, the OpenAI API, and Jira REST APIs that translates natural-language questions into Jira queries and summarizes customer ticket history.
  • Fine-tuned DistilBERT (PyTorch, HuggingFace) to classify ticket intent and route queries to the right tool.
  • Added permission-aware Jira access, input validation, secure config management, API timeouts, and error handling.
PythonFastAPIOpenAI APIJira RESTDistilBERTPyTorch
View on GitHub
ML · Analytics

Customer Churn Prediction & Analytics Dashboard

Classical ML, Explainability & Streamlit

EDA → feature selection → model comparison → SHAP → dashboard

  • Modeled churn on 10,000+ real customer records, comparing Logistic Regression, XGBoost, and LightGBM with stratified cross-validation; evaluated with ROC-AUC and precision/recall and handled class imbalance.
  • Performed exploratory analysis and statistical feature selection in Pandas/Jupyter, explained predictions with SHAP, and tracked experiments in Weights & Biases.
  • Built a Streamlit dashboard with geographic filters and KPI views to present results to non-technical users.
XGBoostLightGBMscikit-learnSHAPW&BPandasStreamlit
View on GitHub
Mobile · Earlier work

Fridgemate

Smart Fridge Companion App

add items → track expiry → reminders

  • A smart fridge companion app that helps you track groceries, reduce food waste, and plan meals.
  • Users add items, set expiry dates, and get reminders before food goes bad.
  • Worked on the front end and project design.
React NativeExpoFirebase
View on GitHub
04

Skills

agentic_aiAgentic AI & LLMs

LangGraphPydantic AILiteLLMRAGSemantic SearchInformation RetrievalTool UseLangfuse

ml_statsMachine Learning & Statistics

PyTorchHuggingFacescikit-learnXGBoostLightGBMPandasNumPyJupyterMatplotlibW&BRegressionCross-validation

data_cloudData, Cloud & MLOps

Postgres/pgvectorSQLiteOpenSearchAWS S3EC2RDSAthenaLambdaAzureDockerFastAPIStreamlitGradioGitMOVEit MFT

languagesLanguages

PythonSQLJavaScriptRustGo
05

Certifications

Cisco

CyberOps Associate

View certificate Stanford & DeepLearning.AI

Machine Learning Specialization

View certificate Anthropic

AI Fluency

View certificate Anthropic

Claude 101

View certificate

Also completed Python for Everybody · Understanding the Brain (University of Chicago)

06 · Contact

Let's build something intelligent.

I'm looking for AI engineering roles — agents, retrieval, evaluation, and the infrastructure around them. The fastest way to reach me is email.

Email suchitpatchigolla@gmail.com LinkedIn in/suchit-raj GitHub Suchit2720