Skip to content
View vishnup22's full-sized avatar
  • United States

Block or report vishnup22

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
vishnup22/README.md

Vishnu

AI Engineer · LLM Systems · AI Research


> whoami

I am an AI Engineer and Researcher working on systems that reason, retrieve, remember, evaluate, and adapt.

My work currently sits across:

LLM Systems       → retrieval, agents, evaluation, post-training
Memory            → persistent context, retrieval, agent memory
Language Models   → multilingual + low-resource NLP
AI Systems        → inference, optimization, reliability
Research          → evaluation, fairness, retrieval stability
Physical AI       → robotics, embodied agents, perception → action

I am especially interested in where these areas begin to converge:

Models × Memory × Agents × Perception × Action

The question behind much of my work is simple:

How do we know an intelligent system actually became better — rather than simply becoming more complex?

That is why I care deeply about evaluation, reproducibility, failure analysis, and measurable system behavior.


🧠 Research Directions

🧠 AI Memory

I am interested in AI systems that can accumulate useful experience instead of starting nearly from zero every interaction.

Areas I want to explore:

  • episodic memory
  • semantic memory
  • working memory
  • memory consolidation
  • retrieval policies
  • memory decay / forgetting
  • shared multi-agent memory
  • long-term personalized agents
  • memory evaluation

Core question

What should an intelligent system remember, retrieve, update, or deliberately forget?

🤖 Physical AI

I am increasingly exploring intelligence beyond text-only environments.

Areas I am interested in:

  • embodied agents
  • robot learning
  • Vision-Language-Action models
  • multimodal foundation models
  • world models
  • robotic planning
  • manipulation
  • reinforcement learning
  • simulation-to-real transfer
  • robot memory

Long-term direction

Connect perception, reasoning, memory, planning, and action inside the same intelligent system.


⚡ Selected AI Systems

🔍 Production RAG Evaluation Platform

An evaluation-first RAG platform designed to detect whether changes to retrieval or generation actually improve the system.

Documents
    │
    ▼
┌───────────────┐
│   Retrieval   │
│  3 strategies │
└───────┬───────┘
        │
        ▼
┌───────────────┐
│     LLM       │
│  3 providers  │
└───────┬───────┘
        │
        ▼
┌─────────────────────────────┐
│         Evaluation          │
│                             │
│ Retrieval Hit Rate          │
│ Faithfulness                │
│ Answer Quality              │
│ Latency                     │
│ Regression Detection        │
└──────────────┬──────────────┘
               │
               ▼
        GitHub Actions Gate

Current evaluation

100% retrieval hit rate · 92% faithfulness

Changes to prompts, retrieval configurations, or model behavior are evaluated through CI regression gates before being accepted.

Treating RAG quality as a software regression problem.


🧩 FinSight AI — Multi-Agent Research System

A four-agent asynchronous research system operating across US, NSE, and BSE markets.

                        ┌────────────────┐
                        │   User Query   │
                        └───────┬────────┘
                                │
               ┌────────────────┼────────────────┐
               │                │                │
               ▼                ▼                ▼
         Research Agent   Sentiment Agent   Market Agent
               │                │                │
               └───────────┬────┴────────────────┘
                           ▼
                    Analysis Agent
                           │
                           ▼
                    LLM Evaluation
                           │
                           ▼
                  Portfolio Optimizer

Combines:

  • asynchronous agent orchestration
  • FinBERT sentiment
  • structured reasoning
  • LLM-as-judge evaluation
  • Markowitz optimization

The research problem I care about here is not simply:

“Can multiple agents talk?”

It is:

How do agent state, memory, handoffs, evaluation, and failure propagation affect the reliability of the final system?


🌐 tok-adapt

A Python library and CLI for adapting pretrained language models to new languages and domains.

Tokenizer
    ↓
Vocabulary Adaptation
    ↓
Embedding Initialization
    ↓
Continued Pretraining
    ↓
Supervised Fine-Tuning
    ↓
DPO

CPT → SFT → DPO

Supports:

Tokenizer Expansion · Vocabulary Replacement · Embedding Adaptation · Weight Initialization · Continued Pretraining · SFT · DPO

The broader research question is how tokenization and representation capacity influence adaptation, especially for languages poorly represented during pretraining.


⚙️ TensorRT Vision Inference

A GPU inference study measuring the effect of TensorRT compilation on ResNet-50 serving performance.

Result
Throughput 97 → 347 FPS
Throughput gain 3.55×
P99 latency 4.3× lower
Accuracy loss 0

Stack:

PyTorch → ONNX → TensorRT 11 → FP16 → CUDA 12.4

The focus is not only optimization.

It is understanding where performance improvements come from and measuring them reproducibly.


📚 Research

🌏 Dravidian Language Models

Evaluating Dedicated Monolingual and Joint Multilingual Causal Models for Dravidian Languages

arXiv Code

Models and experiments across:

తెలుగు Telugu · தமிழ் Tamil · ಕನ್ನಡ Kannada · മലയാളം Malayalam

I study whether dedicated language models can outperform multilingual models when data and compute are constrained.

Evaluation spans:

Language Modeling
├── Loss
├── Perplexity
└── Bits-per-byte

Tokenization
└── Tokenizer efficiency

Downstream Transfer
├── Sentiment
└── Named Entity Recognition

Compared against multilingual baselines including mGPT, mBERT, and XLM-R.


⚡ Reward Model Inference

How Fast Can Reward Models Score?

arXiv Code

A systems study of reward-model inference for RLHF pipelines.

PyTorch Eager
      │
      │ benchmark
      ▼
ONNX Runtime
      │
      ▼
Execution / Graph Analysis

The key finding:

The observed CPU speedup is primarily explained by ONNX Runtime execution and graph optimizations, rather than C++ itself.

Research intersection:

RLHF Infrastructure × Inference × Benchmarking × Systems Analysis


⚖️ Release-Level Fairness Drift

A longitudinal evaluation framework for studying whether model fairness remains stable across releases.

Scope Scale
Model versions 12
Providers 5
Model families 6
Bias benchmarks 7
Evaluation examples ~200K

The central idea:

Fairness should be regression-tested across model releases, not measured once and forgotten.

Status: Under double-blind review.


🔐 Retrieval Stability

Researching certifiable top-k retrieval stability under multi-agent communication.

Datasets:

HotpotQA · 2WikiMultihopQA · MuSiQue

The work studies whether retrieval rankings can remain stable under bounded perturbations introduced during agent communication.

Original Query
      │
      ▼
   Retriever ──────────────► Top-k
      ▲
      │
Agent Communication
      │
      ▼
Perturbed Query
      │
      ▼
   Retriever ──────────────► Top-k'
                              │
                              ▼
                      Stability Certificate

Rather than asking only:

Did retrieval change?

I am interested in:

Can we determine when retrieval is guaranteed not to change?


🧠 Memory × Agents

This is one of the directions I want to push much deeper.

Modern agents can reason over impressive context lengths, but context is not memory.

I am interested in architectures where an agent develops persistent internal state across interactions:

             ┌─────────────┐
             │   Working   │
             │    Memory   │
             └──────┬──────┘
                    │
       ┌────────────┼────────────┐
       ▼            ▼            ▼
   Episodic      Semantic     External
    Memory        Memory       Retrieval
       │            │            │
       └────────────┼────────────┘
                    ▼
              Agent Reasoning
                    │
                    ▼
                  Action
                    │
                    ▼
               Experience
                    │
                    └──────────► Memory Update

Questions I want to investigate:

  • What deserves to become long-term memory?
  • When should memory be retrieved?
  • How should memories decay?
  • How should contradictory memories be resolved?
  • Can agents consolidate multiple experiences into abstractions?
  • How should agent memory be benchmarked?
  • How should memory operate in multi-agent systems?
  • How does memory change long-horizon reasoning?

🤖 Physical AI

The direction I find most exciting is moving intelligence from:

Prompt → Model → Response

toward:

Perception
    ↓
World Understanding
    ↓
Memory
    ↓
Reasoning
    ↓
Planning
    ↓
Action
    ↓
Environment
    └──────────► New Experience

I am especially interested in:

My long-term research interest is in systems that can perceive, reason, remember previous interactions, plan, and act in physical environments.


🧪 Research Philosophy

I like working across the full loop:

BUILD → MEASURE → BREAK → UNDERSTAND → IMPROVE

For me, a model benchmark is rarely the end of the project.

I want to know:

Why did it improve?
Where does it fail?
Does the improvement generalize?
What changed internally?
What happens under distribution shift?
Can the result be reproduced?
Will the system remain reliable after the next change?

🛠 AI Stack

Models & Training

Transformers · Sentence Transformers · SentencePiece
CPT · SFT · DPO


LLM Systems

RAG · Agents · Evaluation · LLM-as-Judge
Retrieval · Structured Generation · Memory


Inference



AI Engineering

MLflow · Prometheus · Evidently


🔭 What I Want to Build Next

                    Intelligent Systems
                           │
          ┌────────────────┼────────────────┐
          │                │                │
          ▼                ▼                ▼
        Models           Memory           Agents
          │                │                │
          └────────────────┼────────────────┘
                           ▼
                       Reasoning
                           │
              ┌────────────┴────────────┐
              ▼                         ▼
          Perception                  Planning
              │                         │
              └────────────┬────────────┘
                           ▼
                         Action
                           │
                           ▼
                     Physical World

I want to work toward AI systems that don't just produce better responses.

I want to understand how to build systems that can:

reason · remember · adapt · learn · perceive · plan · act


Let's Build

I am interested in AI Engineer / Research Engineer opportunities and research collaborations around:

LLM Systems · Agents · Memory · Evaluation · Multilingual NLP · Efficient Inference · Physical AI · Robotics




Pinned Loading

  1. FinSight-AI FinSight-AI Public

    Python

  2. flash-sale-reservation-platform flash-sale-reservation-platform Public

    High-throughput flash-sale reservation platform with Redis/Redlock inventory holds, Kafka-backed checkout processing, PostgreSQL consistency, and real- time seat availability.

    Java

  3. flowforge-automation-engine flowforge-automation-engine Public

    A visual workflow automation engine with a React Flow editor, Spring Boot DAG execution, real-time SSE logs, and encrypted OAuth credentials.

    Java

  4. production-rag-eval-platform production-rag-eval-platform Public

    RAG evaluation platform with CI/CD quality gates that blocks bad prompt, model, and retriever changes before deployment.

    Python

  5. quire-workspace quire-workspace Public

    Collaborative workspace with live CRDT editing — React/Tiptap/Yjs + Spring Boot/Postgres/Redis

    Java

  6. tensorrt-vision-server tensorrt-vision-server Public

    ResNet-50 served via TensorRT FP16 + C++ Crow HTTP server. 3.55x FPS, 4.3x P99 latency reduction, zero accuracy loss vs PyTorch baseline.

    Python