Monesh Rallapalli.

Available · AI Engineer

MoneshRallapalli.

I build AI systems that see, retrieve, reason & ship.

CS master's graduate from Indiana University focused on agentic pipelines, multimodal retrieval, production ML, and cloud-backed systems that hold up outside the notebook.

Based
Bloomington, Indiana
Focus
Agentic & multimodal systems
Status
Open to full-time roles

Case studies in retrieval, agents & distributed systems.

§ 01 · Work
01 In progress · Building now

ava_companion

A LangGraph-powered AI companion — text-first, memory-aware, and built to feel like texting a real person. It's live and still taking shape, which is exactly why it's here early. Tell me what it should do next.

Shipping now
  • Intelligent conversation — Groq-powered chat with a system prompt driven by live context (activity + memories).
  • Smart routing — classifies each message as text, image, or audio and hands it to the right workflow.
  • “What am I doing right now?” context — schedule-based awareness so replies match Ava's current activity.
  • Long-term memory — extracts facts from your messages, stores them in Qdrant, and injects relevant ones into future replies.
  • LangGraph agent pipeline — the full graph wired in production order: extract → route → context → inject → respond.
Coming soon
  • Voice in/out (Whisper + ElevenLabs)
  • Image generation & vision (FLUX + Llama Vision)
  • Chainlit demo UI
  • WhatsApp integration
  • Cloud deployment

Full write-up lands here once it ships.

Suggest a direction

Fig. 2 · System architecture2026
02Search over long-form video

Multimodal Video RAG Platform

A retrieval system that answers questions over visual frames and spoken transcript, then returns timestamped evidence you can jump back to.

0%
Retrieval recall
0%
Grounded answers
0%
No-answer F1
0
Hand-labeled queries
  • LangGraph
  • FastAPI
  • AWS
  • Pinecone
  • Bedrock
  • Next.js
Fig. 3 · System architecture2026
03Real-time multi-agent surveillance

AI-Powered Video Intelligence Platform

A multi-agent surveillance system: one agent watches every frame, one reasons about what's worth an alert, one turns plain-English commands into monitoring rules, and one keeps searchable memory of everything it has seen.

0
Cooperating AI agents
0
AI providers
0
Realtime WS channels
0
Alert severity tiers
  • Multi-Agent AI
  • Claude Sonnet
  • Gemini
  • FastAPI
  • WebSocket
  • ChromaDB
  • PostgreSQL
Fig. 4 · System architecture2025
04Event-driven booking at scale

Room Reservation Platform

A campus room-booking platform where students, faculty, and staff browse an interactive floor plan and reserve rooms by role, with per-role limits enforced at the API gateway. An event-driven microservice backend communicates asynchronously over RabbitMQ so the booking flow stays resilient under load.

0
Req/s · k6 stress test
1.6M
Requests load-tested
1.06ms
Avg response
3.67ms
p95 latency
  • Spring Boot
  • Angular
  • PostgreSQL
  • Redis
  • RabbitMQ
  • Kubernetes
  • k6
Fig. 5 · System architecture2025
05Multi-LLM travel planner

No Detours

An LLM travel planner that turns a plain-English request into a day-by-day itinerary, packing list, and budget. A guardrail-fronted agent extracts trip features, fans out to five live data sources in parallel, then composes the plan through a provider-agnostic LLM layer with automatic fallback — backed by a judge-LLM harness that scores plan quality across five metrics.

0%
Faster via parallel async
5+
Live data sources fused
0
LLM-judged metrics
0
Delivery modes · web · API · CLI
  • Python
  • FastAPI
  • LangChain
  • OpenAI
  • Anthropic Claude
  • asyncio

A path shaped by AI, systems & applied research.

§ 02 · Career

Experience

Machine Learning Intern

MyEdMaster LLC, Leesburg, VA

  • Built a real-time posture assessment system for 3 exercise types using MediaPipe joint-angle extraction
  • Containerized ML services with Docker and automated AWS CI/CD using GitHub Actions

Machine Learning Engineer Intern

Fox Trading (1STOP.AI), Remote

  • Built an end-to-end diabetes-prediction model on structured healthcare data using supervised learning
  • Ran data preprocessing and feature engineering, then benchmarked classification algorithms for accuracy

Machine Learning Engineer Intern

Infosol Technosol, Remote

  • Built a customer-churn framework across 5 models, reaching 0.845 ROC-AUC with MLflow experiment tracking
  • Lifted XGBoost precision 16% via SMOTE balancing and cut the feature space 29% (34 → 24 features)

Education

MS Computer Science

Indiana University Bloomington

GPA 3.80

BTech CS · AI Specialization

Jain University

GPA 8.97 / 10

The tools I reach for when a prototype must become a system.

§ 03 · Stack

A. AI / ML

LangChainLangGraphRAGNLPMulti-AgentTensorFlowPyTorch

B. Backend

FastAPISpring BootCeleryPostgreSQLRedisRabbitMQ

C. Cloud

AWSGCPDockerKubernetesCI/CD

D. Search

PineconeChromaDBBM25Vector DBs

The slower work of formalizing ideas.

§ 04 · Research
[2]

Haptic Feedback-Enabled Writing Skills Improvement Device

Patent Office Journal · 2024

Patent

§ 05 · Contact

Let's build something that ships.

Hiring, have a lead, or want to talk agentic systems, multimodal RAG, or production ML? My inbox is open.

Send a message